Hugging Face's logo

SGPatil
/
GR00T-N1.5-SO101-finetune-table-cleanup

Safetensors
gr00t_n1_5
Model card Files
xet
Community
SGPatil commited on
Commit
6858f84
·
verified ·
1 Parent(s): 65f94a7

Upload folder using huggingface_hub

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. checkpoint-10000/config.json +64 -0
  2. checkpoint-10000/experiment_cfg/metadata.json +195 -0
  3. checkpoint-10000/model-00001-of-00002.safetensors +3 -0
  4. checkpoint-10000/model-00002-of-00002.safetensors +3 -0
  5. checkpoint-10000/model.safetensors.index.json +0 -0
  6. checkpoint-10000/optimizer.pt +3 -0
  7. checkpoint-10000/rng_state.pth +3 -0
  8. checkpoint-10000/scheduler.pt +3 -0
  9. checkpoint-10000/trainer_state.json +0 -0
  10. checkpoint-4000/config.json +64 -0
  11. checkpoint-4000/experiment_cfg/metadata.json +195 -0
  12. checkpoint-4000/model-00001-of-00002.safetensors +3 -0
  13. checkpoint-4000/model-00002-of-00002.safetensors +3 -0
  14. checkpoint-4000/model.safetensors.index.json +0 -0
  15. checkpoint-4000/optimizer.pt +3 -0
  16. checkpoint-4000/rng_state.pth +3 -0
  17. checkpoint-4000/scheduler.pt +3 -0
  18. checkpoint-4000/trainer_state.json +2834 -0
  19. checkpoint-5000/config.json +64 -0
  20. checkpoint-5000/experiment_cfg/metadata.json +195 -0
  21. checkpoint-5000/model-00001-of-00002.safetensors +3 -0
  22. checkpoint-5000/model-00002-of-00002.safetensors +3 -0
  23. checkpoint-5000/model.safetensors.index.json +0 -0
  24. checkpoint-5000/optimizer.pt +3 -0
  25. checkpoint-5000/rng_state.pth +3 -0
  26. checkpoint-5000/scheduler.pt +3 -0
  27. checkpoint-5000/trainer_state.json +3534 -0
  28. checkpoint-6000/config.json +64 -0
  29. checkpoint-6000/experiment_cfg/metadata.json +195 -0
  30. checkpoint-6000/model-00001-of-00002.safetensors +3 -0
  31. checkpoint-6000/model-00002-of-00002.safetensors +3 -0
  32. checkpoint-6000/model.safetensors.index.json +0 -0
  33. checkpoint-6000/optimizer.pt +3 -0
  34. checkpoint-6000/rng_state.pth +3 -0
  35. checkpoint-6000/scheduler.pt +3 -0
  36. checkpoint-6000/trainer_state.json +0 -0
  37. checkpoint-7000/config.json +64 -0
  38. checkpoint-7000/experiment_cfg/metadata.json +195 -0
  39. checkpoint-7000/model-00001-of-00002.safetensors +3 -0
  40. checkpoint-7000/model-00002-of-00002.safetensors +3 -0
  41. checkpoint-7000/model.safetensors.index.json +0 -0
  42. checkpoint-7000/optimizer.pt +3 -0
  43. checkpoint-7000/rng_state.pth +3 -0
  44. checkpoint-7000/scheduler.pt +3 -0
  45. checkpoint-7000/trainer_state.json +0 -0
  46. checkpoint-8000/config.json +64 -0
  47. checkpoint-8000/experiment_cfg/metadata.json +195 -0
  48. checkpoint-8000/model-00001-of-00002.safetensors +3 -0
  49. checkpoint-8000/model-00002-of-00002.safetensors +3 -0
  50. checkpoint-8000/model.safetensors.index.json +0 -0
checkpoint-10000/config.json ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "action_dim": 32,
3
+ "action_head_cfg": {
4
+ "action_dim": 32,
5
+ "action_horizon": 16,
6
+ "add_pos_embed": true,
7
+ "backbone_embedding_dim": 2048,
8
+ "diffusion_model_cfg": {
9
+ "attention_head_dim": 48,
10
+ "cross_attention_dim": 2048,
11
+ "dropout": 0.2,
12
+ "final_dropout": true,
13
+ "interleave_self_attention": true,
14
+ "norm_type": "ada_norm",
15
+ "num_attention_heads": 32,
16
+ "num_layers": 16,
17
+ "output_dim": 1024,
18
+ "positional_embeddings": null
19
+ },
20
+ "hidden_size": 1024,
21
+ "input_embedding_dim": 1536,
22
+ "max_action_dim": 32,
23
+ "max_state_dim": 64,
24
+ "model_dtype": "float32",
25
+ "noise_beta_alpha": 1.5,
26
+ "noise_beta_beta": 1.0,
27
+ "noise_s": 0.999,
28
+ "num_inference_timesteps": 4,
29
+ "num_target_vision_tokens": 32,
30
+ "num_timestep_buckets": 1000,
31
+ "tune_diffusion_model": true,
32
+ "tune_projector": true,
33
+ "use_vlln": true,
34
+ "vl_self_attention_cfg": {
35
+ "attention_head_dim": 64,
36
+ "dropout": 0.2,
37
+ "final_dropout": true,
38
+ "num_attention_heads": 32,
39
+ "num_layers": 4,
40
+ "positional_embeddings": null
41
+ }
42
+ },
43
+ "action_horizon": 16,
44
+ "architectures": [
45
+ "GR00T_N1_5"
46
+ ],
47
+ "attn_implementation": null,
48
+ "backbone_cfg": {
49
+ "eagle_path": "NVEagle/eagle_er-qwen3_1_7B-Siglip2_400M_stage1_5_128gpu_er_v7_1mlp_nops",
50
+ "load_bf16": false,
51
+ "project_to_dim": null,
52
+ "reproject_vision": false,
53
+ "select_layer": 12,
54
+ "tune_llm": false,
55
+ "tune_visual": true,
56
+ "use_flash_attention": true
57
+ },
58
+ "compute_dtype": "bfloat16",
59
+ "hidden_size": 2048,
60
+ "model_dtype": "float32",
61
+ "model_type": "gr00t_n1_5",
62
+ "torch_dtype": "bfloat16",
63
+ "transformers_version": "4.51.3"
64
+ }
checkpoint-10000/experiment_cfg/metadata.json ADDED
@@ -0,0 +1,195 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "new_embodiment": {
3
+ "statistics": {
4
+ "state": {
5
+ "single_arm": {
6
+ "max": [
7
+ 33.22265625,
8
+ 86.1328125,
9
+ 84.7265625,
10
+ 101.337890625,
11
+ 32.783203125
12
+ ],
13
+ "min": [
14
+ -87.275390625,
15
+ -103.974609375,
16
+ -106.435546875,
17
+ -16.34765625,
18
+ -170.068359375
19
+ ],
20
+ "mean": [
21
+ -26.84341812133789,
22
+ 6.497307777404785,
23
+ -22.14896583557129,
24
+ 74.85757446289062,
25
+ -111.02351379394531
26
+ ],
27
+ "std": [
28
+ 27.50257110595703,
29
+ 41.151065826416016,
30
+ 48.41417694091797,
31
+ 20.208982467651367,
32
+ 46.13625717163086
33
+ ],
34
+ "q01": [
35
+ -79.013671875,
36
+ -103.271484375,
37
+ -104.0625,
38
+ 19.187402343750005,
39
+ -169.892578125
40
+ ],
41
+ "q99": [
42
+ 24.169921875,
43
+ 73.30078125,
44
+ 84.375,
45
+ 101.07421875,
46
+ 11.162109375
47
+ ]
48
+ },
49
+ "gripper": {
50
+ "max": [
51
+ 92.373046875
52
+ ],
53
+ "min": [
54
+ -0.615234375
55
+ ],
56
+ "mean": [
57
+ 14.521344184875488
58
+ ],
59
+ "std": [
60
+ 13.066176414489746
61
+ ],
62
+ "q01": [
63
+ 0.0
64
+ ],
65
+ "q99": [
66
+ 52.470703125
67
+ ]
68
+ }
69
+ },
70
+ "action": {
71
+ "single_arm": {
72
+ "max": [
73
+ 33.662109375,
74
+ 86.484375,
75
+ 85.078125,
76
+ 102.65625,
77
+ 33.3984375
78
+ ],
79
+ "min": [
80
+ -87.890625,
81
+ -104.4140625,
82
+ -110.7421875,
83
+ -19.599609375,
84
+ -173.3203125
85
+ ],
86
+ "mean": [
87
+ -26.84657096862793,
88
+ 4.594353199005127,
89
+ -23.747364044189453,
90
+ 74.75859832763672,
91
+ -111.30194854736328
92
+ ],
93
+ "std": [
94
+ 27.627857208251953,
95
+ 40.672019958496094,
96
+ 48.966896057128906,
97
+ 20.728727340698242,
98
+ 46.57689666748047
99
+ ],
100
+ "q01": [
101
+ -79.453125,
102
+ -102.744140625,
103
+ -109.6875,
104
+ 17.9296875,
105
+ -172.177734375
106
+ ],
107
+ "q99": [
108
+ 24.08203125,
109
+ 72.685546875,
110
+ 83.84765625,
111
+ 101.77734375,
112
+ 10.72265625
113
+ ]
114
+ },
115
+ "gripper": {
116
+ "max": [
117
+ 92.63671875
118
+ ],
119
+ "min": [
120
+ -3.076171875
121
+ ],
122
+ "mean": [
123
+ 12.46182632446289
124
+ ],
125
+ "std": [
126
+ 14.029601097106934
127
+ ],
128
+ "q01": [
129
+ -0.52734375
130
+ ],
131
+ "q99": [
132
+ 52.91015625
133
+ ]
134
+ }
135
+ }
136
+ },
137
+ "modalities": {
138
+ "video": {
139
+ "front": {
140
+ "resolution": [
141
+ 640,
142
+ 480
143
+ ],
144
+ "channels": 3,
145
+ "fps": 30.0
146
+ },
147
+ "laptop": {
148
+ "resolution": [
149
+ 640,
150
+ 480
151
+ ],
152
+ "channels": 3,
153
+ "fps": 30.0
154
+ }
155
+ },
156
+ "state": {
157
+ "single_arm": {
158
+ "absolute": true,
159
+ "rotation_type": null,
160
+ "shape": [
161
+ 5
162
+ ],
163
+ "continuous": true
164
+ },
165
+ "gripper": {
166
+ "absolute": true,
167
+ "rotation_type": null,
168
+ "shape": [
169
+ 1
170
+ ],
171
+ "continuous": true
172
+ }
173
+ },
174
+ "action": {
175
+ "single_arm": {
176
+ "absolute": true,
177
+ "rotation_type": null,
178
+ "shape": [
179
+ 5
180
+ ],
181
+ "continuous": true
182
+ },
183
+ "gripper": {
184
+ "absolute": true,
185
+ "rotation_type": null,
186
+ "shape": [
187
+ 1
188
+ ],
189
+ "continuous": true
190
+ }
191
+ }
192
+ },
193
+ "embodiment_tag": "new_embodiment"
194
+ }
195
+ }
checkpoint-10000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0e65895a7b6c052e58b679bdd68bff54a68694a81f38e23d3e03a92059b3600b
3
+ size 4999367032
checkpoint-10000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:595f7bca8c51123b14908af4b1eac93a96d52a59a6b0fe7d8bae323aa957c147
3
+ size 2586705312
checkpoint-10000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-10000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:da86f5adf2d73c513c74e0a03ac6ef1580f785e88602acbe8e24e4c6d9e311d3
3
+ size 4147424698
checkpoint-10000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:19496a2c7f714f9d36473c01b7fd34ce69b238124e9968608454712b8e5755ba
3
+ size 14244
checkpoint-10000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:07356cb668b8efc440ddb2f7f81f7da3152060e81d0229922fa3b30df3313f82
3
+ size 1064
checkpoint-10000/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-4000/config.json ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "action_dim": 32,
3
+ "action_head_cfg": {
4
+ "action_dim": 32,
5
+ "action_horizon": 16,
6
+ "add_pos_embed": true,
7
+ "backbone_embedding_dim": 2048,
8
+ "diffusion_model_cfg": {
9
+ "attention_head_dim": 48,
10
+ "cross_attention_dim": 2048,
11
+ "dropout": 0.2,
12
+ "final_dropout": true,
13
+ "interleave_self_attention": true,
14
+ "norm_type": "ada_norm",
15
+ "num_attention_heads": 32,
16
+ "num_layers": 16,
17
+ "output_dim": 1024,
18
+ "positional_embeddings": null
19
+ },
20
+ "hidden_size": 1024,
21
+ "input_embedding_dim": 1536,
22
+ "max_action_dim": 32,
23
+ "max_state_dim": 64,
24
+ "model_dtype": "float32",
25
+ "noise_beta_alpha": 1.5,
26
+ "noise_beta_beta": 1.0,
27
+ "noise_s": 0.999,
28
+ "num_inference_timesteps": 4,
29
+ "num_target_vision_tokens": 32,
30
+ "num_timestep_buckets": 1000,
31
+ "tune_diffusion_model": true,
32
+ "tune_projector": true,
33
+ "use_vlln": true,
34
+ "vl_self_attention_cfg": {
35
+ "attention_head_dim": 64,
36
+ "dropout": 0.2,
37
+ "final_dropout": true,
38
+ "num_attention_heads": 32,
39
+ "num_layers": 4,
40
+ "positional_embeddings": null
41
+ }
42
+ },
43
+ "action_horizon": 16,
44
+ "architectures": [
45
+ "GR00T_N1_5"
46
+ ],
47
+ "attn_implementation": null,
48
+ "backbone_cfg": {
49
+ "eagle_path": "NVEagle/eagle_er-qwen3_1_7B-Siglip2_400M_stage1_5_128gpu_er_v7_1mlp_nops",
50
+ "load_bf16": false,
51
+ "project_to_dim": null,
52
+ "reproject_vision": false,
53
+ "select_layer": 12,
54
+ "tune_llm": false,
55
+ "tune_visual": true,
56
+ "use_flash_attention": true
57
+ },
58
+ "compute_dtype": "bfloat16",
59
+ "hidden_size": 2048,
60
+ "model_dtype": "float32",
61
+ "model_type": "gr00t_n1_5",
62
+ "torch_dtype": "bfloat16",
63
+ "transformers_version": "4.51.3"
64
+ }
checkpoint-4000/experiment_cfg/metadata.json ADDED
@@ -0,0 +1,195 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "new_embodiment": {
3
+ "statistics": {
4
+ "state": {
5
+ "single_arm": {
6
+ "max": [
7
+ 33.22265625,
8
+ 86.1328125,
9
+ 84.7265625,
10
+ 101.337890625,
11
+ 32.783203125
12
+ ],
13
+ "min": [
14
+ -87.275390625,
15
+ -103.974609375,
16
+ -106.435546875,
17
+ -16.34765625,
18
+ -170.068359375
19
+ ],
20
+ "mean": [
21
+ -26.84341812133789,
22
+ 6.497307777404785,
23
+ -22.14896583557129,
24
+ 74.85757446289062,
25
+ -111.02351379394531
26
+ ],
27
+ "std": [
28
+ 27.50257110595703,
29
+ 41.151065826416016,
30
+ 48.41417694091797,
31
+ 20.208982467651367,
32
+ 46.13625717163086
33
+ ],
34
+ "q01": [
35
+ -79.013671875,
36
+ -103.271484375,
37
+ -104.0625,
38
+ 19.187402343750005,
39
+ -169.892578125
40
+ ],
41
+ "q99": [
42
+ 24.169921875,
43
+ 73.30078125,
44
+ 84.375,
45
+ 101.07421875,
46
+ 11.162109375
47
+ ]
48
+ },
49
+ "gripper": {
50
+ "max": [
51
+ 92.373046875
52
+ ],
53
+ "min": [
54
+ -0.615234375
55
+ ],
56
+ "mean": [
57
+ 14.521344184875488
58
+ ],
59
+ "std": [
60
+ 13.066176414489746
61
+ ],
62
+ "q01": [
63
+ 0.0
64
+ ],
65
+ "q99": [
66
+ 52.470703125
67
+ ]
68
+ }
69
+ },
70
+ "action": {
71
+ "single_arm": {
72
+ "max": [
73
+ 33.662109375,
74
+ 86.484375,
75
+ 85.078125,
76
+ 102.65625,
77
+ 33.3984375
78
+ ],
79
+ "min": [
80
+ -87.890625,
81
+ -104.4140625,
82
+ -110.7421875,
83
+ -19.599609375,
84
+ -173.3203125
85
+ ],
86
+ "mean": [
87
+ -26.84657096862793,
88
+ 4.594353199005127,
89
+ -23.747364044189453,
90
+ 74.75859832763672,
91
+ -111.30194854736328
92
+ ],
93
+ "std": [
94
+ 27.627857208251953,
95
+ 40.672019958496094,
96
+ 48.966896057128906,
97
+ 20.728727340698242,
98
+ 46.57689666748047
99
+ ],
100
+ "q01": [
101
+ -79.453125,
102
+ -102.744140625,
103
+ -109.6875,
104
+ 17.9296875,
105
+ -172.177734375
106
+ ],
107
+ "q99": [
108
+ 24.08203125,
109
+ 72.685546875,
110
+ 83.84765625,
111
+ 101.77734375,
112
+ 10.72265625
113
+ ]
114
+ },
115
+ "gripper": {
116
+ "max": [
117
+ 92.63671875
118
+ ],
119
+ "min": [
120
+ -3.076171875
121
+ ],
122
+ "mean": [
123
+ 12.46182632446289
124
+ ],
125
+ "std": [
126
+ 14.029601097106934
127
+ ],
128
+ "q01": [
129
+ -0.52734375
130
+ ],
131
+ "q99": [
132
+ 52.91015625
133
+ ]
134
+ }
135
+ }
136
+ },
137
+ "modalities": {
138
+ "video": {
139
+ "front": {
140
+ "resolution": [
141
+ 640,
142
+ 480
143
+ ],
144
+ "channels": 3,
145
+ "fps": 30.0
146
+ },
147
+ "laptop": {
148
+ "resolution": [
149
+ 640,
150
+ 480
151
+ ],
152
+ "channels": 3,
153
+ "fps": 30.0
154
+ }
155
+ },
156
+ "state": {
157
+ "single_arm": {
158
+ "absolute": true,
159
+ "rotation_type": null,
160
+ "shape": [
161
+ 5
162
+ ],
163
+ "continuous": true
164
+ },
165
+ "gripper": {
166
+ "absolute": true,
167
+ "rotation_type": null,
168
+ "shape": [
169
+ 1
170
+ ],
171
+ "continuous": true
172
+ }
173
+ },
174
+ "action": {
175
+ "single_arm": {
176
+ "absolute": true,
177
+ "rotation_type": null,
178
+ "shape": [
179
+ 5
180
+ ],
181
+ "continuous": true
182
+ },
183
+ "gripper": {
184
+ "absolute": true,
185
+ "rotation_type": null,
186
+ "shape": [
187
+ 1
188
+ ],
189
+ "continuous": true
190
+ }
191
+ }
192
+ },
193
+ "embodiment_tag": "new_embodiment"
194
+ }
195
+ }
checkpoint-4000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0e65895a7b6c052e58b679bdd68bff54a68694a81f38e23d3e03a92059b3600b
3
+ size 4999367032
checkpoint-4000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f9bf9ecec75543ce10ef434c434df87246ed1520853803dbde888a1e053e80b
3
+ size 2586705312
checkpoint-4000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-4000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:39008997be56844954a656172d474c0ac07a1aff1983543078e43e76e3c07cea
3
+ size 4147424698
checkpoint-4000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1b853fa74473f32e44be47a34bdeaf5dc696917c249e44f780bcd71368abd20c
3
+ size 14244
checkpoint-4000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7211b192985f3cf213c0fe5d877c55ab1774647baa7fcf0d4d0c00aad611b2fd
3
+ size 1064
checkpoint-4000/trainer_state.json ADDED
@@ -0,0 +1,2834 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 1.5929908403026682,
6
+ "eval_steps": 500,
7
+ "global_step": 4000,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.00398247710075667,
14
+ "grad_norm": 10.328946113586426,
15
+ "learning_rate": 1.8e-06,
16
+ "loss": 0.9415,
17
+ "step": 10
18
+ },
19
+ {
20
+ "epoch": 0.00796495420151334,
21
+ "grad_norm": 6.881450176239014,
22
+ "learning_rate": 3.8e-06,
23
+ "loss": 0.8639,
24
+ "step": 20
25
+ },
26
+ {
27
+ "epoch": 0.011947431302270013,
28
+ "grad_norm": 2.251582384109497,
29
+ "learning_rate": 5.8e-06,
30
+ "loss": 0.4771,
31
+ "step": 30
32
+ },
33
+ {
34
+ "epoch": 0.01592990840302668,
35
+ "grad_norm": 1.6325163841247559,
36
+ "learning_rate": 7.8e-06,
37
+ "loss": 0.2953,
38
+ "step": 40
39
+ },
40
+ {
41
+ "epoch": 0.019912385503783353,
42
+ "grad_norm": 0.9643872976303101,
43
+ "learning_rate": 9.800000000000001e-06,
44
+ "loss": 0.2403,
45
+ "step": 50
46
+ },
47
+ {
48
+ "epoch": 0.023894862604540025,
49
+ "grad_norm": 1.4014142751693726,
50
+ "learning_rate": 1.18e-05,
51
+ "loss": 0.2213,
52
+ "step": 60
53
+ },
54
+ {
55
+ "epoch": 0.027877339705296694,
56
+ "grad_norm": 1.3301987648010254,
57
+ "learning_rate": 1.3800000000000002e-05,
58
+ "loss": 0.2048,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.03185981680605336,
63
+ "grad_norm": 1.5023683309555054,
64
+ "learning_rate": 1.58e-05,
65
+ "loss": 0.1658,
66
+ "step": 80
67
+ },
68
+ {
69
+ "epoch": 0.035842293906810034,
70
+ "grad_norm": 1.7575840950012207,
71
+ "learning_rate": 1.78e-05,
72
+ "loss": 0.146,
73
+ "step": 90
74
+ },
75
+ {
76
+ "epoch": 0.039824771007566706,
77
+ "grad_norm": 1.4768755435943604,
78
+ "learning_rate": 1.9800000000000004e-05,
79
+ "loss": 0.154,
80
+ "step": 100
81
+ },
82
+ {
83
+ "epoch": 0.04380724810832338,
84
+ "grad_norm": 1.43697988986969,
85
+ "learning_rate": 2.18e-05,
86
+ "loss": 0.1419,
87
+ "step": 110
88
+ },
89
+ {
90
+ "epoch": 0.04778972520908005,
91
+ "grad_norm": 1.6363322734832764,
92
+ "learning_rate": 2.38e-05,
93
+ "loss": 0.1243,
94
+ "step": 120
95
+ },
96
+ {
97
+ "epoch": 0.051772202309836715,
98
+ "grad_norm": 1.29106605052948,
99
+ "learning_rate": 2.58e-05,
100
+ "loss": 0.1273,
101
+ "step": 130
102
+ },
103
+ {
104
+ "epoch": 0.05575467941059339,
105
+ "grad_norm": 0.849351704120636,
106
+ "learning_rate": 2.7800000000000005e-05,
107
+ "loss": 0.1199,
108
+ "step": 140
109
+ },
110
+ {
111
+ "epoch": 0.05973715651135006,
112
+ "grad_norm": 1.442665934562683,
113
+ "learning_rate": 2.98e-05,
114
+ "loss": 0.1068,
115
+ "step": 150
116
+ },
117
+ {
118
+ "epoch": 0.06371963361210672,
119
+ "grad_norm": 1.277969241142273,
120
+ "learning_rate": 3.18e-05,
121
+ "loss": 0.11,
122
+ "step": 160
123
+ },
124
+ {
125
+ "epoch": 0.0677021107128634,
126
+ "grad_norm": 1.636638879776001,
127
+ "learning_rate": 3.38e-05,
128
+ "loss": 0.1001,
129
+ "step": 170
130
+ },
131
+ {
132
+ "epoch": 0.07168458781362007,
133
+ "grad_norm": 1.0019601583480835,
134
+ "learning_rate": 3.58e-05,
135
+ "loss": 0.1126,
136
+ "step": 180
137
+ },
138
+ {
139
+ "epoch": 0.07566706491437675,
140
+ "grad_norm": 0.9693198204040527,
141
+ "learning_rate": 3.7800000000000004e-05,
142
+ "loss": 0.0867,
143
+ "step": 190
144
+ },
145
+ {
146
+ "epoch": 0.07964954201513341,
147
+ "grad_norm": 0.7893789410591125,
148
+ "learning_rate": 3.9800000000000005e-05,
149
+ "loss": 0.1002,
150
+ "step": 200
151
+ },
152
+ {
153
+ "epoch": 0.08363201911589008,
154
+ "grad_norm": 1.736881136894226,
155
+ "learning_rate": 4.18e-05,
156
+ "loss": 0.1015,
157
+ "step": 210
158
+ },
159
+ {
160
+ "epoch": 0.08761449621664676,
161
+ "grad_norm": 0.7016027569770813,
162
+ "learning_rate": 4.38e-05,
163
+ "loss": 0.0895,
164
+ "step": 220
165
+ },
166
+ {
167
+ "epoch": 0.09159697331740342,
168
+ "grad_norm": 1.6043905019760132,
169
+ "learning_rate": 4.58e-05,
170
+ "loss": 0.0857,
171
+ "step": 230
172
+ },
173
+ {
174
+ "epoch": 0.0955794504181601,
175
+ "grad_norm": 1.4260313510894775,
176
+ "learning_rate": 4.78e-05,
177
+ "loss": 0.0818,
178
+ "step": 240
179
+ },
180
+ {
181
+ "epoch": 0.09956192751891677,
182
+ "grad_norm": 0.9992408156394958,
183
+ "learning_rate": 4.9800000000000004e-05,
184
+ "loss": 0.0763,
185
+ "step": 250
186
+ },
187
+ {
188
+ "epoch": 0.10354440461967343,
189
+ "grad_norm": 0.5259194374084473,
190
+ "learning_rate": 5.1800000000000005e-05,
191
+ "loss": 0.0728,
192
+ "step": 260
193
+ },
194
+ {
195
+ "epoch": 0.10752688172043011,
196
+ "grad_norm": 0.8904138207435608,
197
+ "learning_rate": 5.380000000000001e-05,
198
+ "loss": 0.0741,
199
+ "step": 270
200
+ },
201
+ {
202
+ "epoch": 0.11150935882118677,
203
+ "grad_norm": 0.8944669961929321,
204
+ "learning_rate": 5.580000000000001e-05,
205
+ "loss": 0.0646,
206
+ "step": 280
207
+ },
208
+ {
209
+ "epoch": 0.11549183592194345,
210
+ "grad_norm": 1.0091243982315063,
211
+ "learning_rate": 5.7799999999999995e-05,
212
+ "loss": 0.07,
213
+ "step": 290
214
+ },
215
+ {
216
+ "epoch": 0.11947431302270012,
217
+ "grad_norm": 0.7492197155952454,
218
+ "learning_rate": 5.9800000000000003e-05,
219
+ "loss": 0.0726,
220
+ "step": 300
221
+ },
222
+ {
223
+ "epoch": 0.12345679012345678,
224
+ "grad_norm": 0.9669196605682373,
225
+ "learning_rate": 6.18e-05,
226
+ "loss": 0.068,
227
+ "step": 310
228
+ },
229
+ {
230
+ "epoch": 0.12743926722421345,
231
+ "grad_norm": 0.8658894896507263,
232
+ "learning_rate": 6.38e-05,
233
+ "loss": 0.0605,
234
+ "step": 320
235
+ },
236
+ {
237
+ "epoch": 0.13142174432497014,
238
+ "grad_norm": 0.6594343185424805,
239
+ "learning_rate": 6.58e-05,
240
+ "loss": 0.06,
241
+ "step": 330
242
+ },
243
+ {
244
+ "epoch": 0.1354042214257268,
245
+ "grad_norm": 0.9863249659538269,
246
+ "learning_rate": 6.780000000000001e-05,
247
+ "loss": 0.0582,
248
+ "step": 340
249
+ },
250
+ {
251
+ "epoch": 0.13938669852648347,
252
+ "grad_norm": 0.8012584447860718,
253
+ "learning_rate": 6.98e-05,
254
+ "loss": 0.0581,
255
+ "step": 350
256
+ },
257
+ {
258
+ "epoch": 0.14336917562724014,
259
+ "grad_norm": 0.8694135546684265,
260
+ "learning_rate": 7.18e-05,
261
+ "loss": 0.06,
262
+ "step": 360
263
+ },
264
+ {
265
+ "epoch": 0.1473516527279968,
266
+ "grad_norm": 0.8341110944747925,
267
+ "learning_rate": 7.38e-05,
268
+ "loss": 0.0551,
269
+ "step": 370
270
+ },
271
+ {
272
+ "epoch": 0.1513341298287535,
273
+ "grad_norm": 0.8546151518821716,
274
+ "learning_rate": 7.58e-05,
275
+ "loss": 0.0504,
276
+ "step": 380
277
+ },
278
+ {
279
+ "epoch": 0.15531660692951016,
280
+ "grad_norm": 0.8565087914466858,
281
+ "learning_rate": 7.780000000000001e-05,
282
+ "loss": 0.0544,
283
+ "step": 390
284
+ },
285
+ {
286
+ "epoch": 0.15929908403026682,
287
+ "grad_norm": 0.9835138320922852,
288
+ "learning_rate": 7.98e-05,
289
+ "loss": 0.0528,
290
+ "step": 400
291
+ },
292
+ {
293
+ "epoch": 0.1632815611310235,
294
+ "grad_norm": 0.9852800965309143,
295
+ "learning_rate": 8.18e-05,
296
+ "loss": 0.0556,
297
+ "step": 410
298
+ },
299
+ {
300
+ "epoch": 0.16726403823178015,
301
+ "grad_norm": 0.9271975755691528,
302
+ "learning_rate": 8.38e-05,
303
+ "loss": 0.053,
304
+ "step": 420
305
+ },
306
+ {
307
+ "epoch": 0.17124651533253685,
308
+ "grad_norm": 0.843838632106781,
309
+ "learning_rate": 8.58e-05,
310
+ "loss": 0.0536,
311
+ "step": 430
312
+ },
313
+ {
314
+ "epoch": 0.1752289924332935,
315
+ "grad_norm": 0.6337900161743164,
316
+ "learning_rate": 8.78e-05,
317
+ "loss": 0.0501,
318
+ "step": 440
319
+ },
320
+ {
321
+ "epoch": 0.17921146953405018,
322
+ "grad_norm": 0.8339642882347107,
323
+ "learning_rate": 8.98e-05,
324
+ "loss": 0.0535,
325
+ "step": 450
326
+ },
327
+ {
328
+ "epoch": 0.18319394663480684,
329
+ "grad_norm": 0.8911416530609131,
330
+ "learning_rate": 9.180000000000001e-05,
331
+ "loss": 0.0599,
332
+ "step": 460
333
+ },
334
+ {
335
+ "epoch": 0.1871764237355635,
336
+ "grad_norm": 0.7793362736701965,
337
+ "learning_rate": 9.38e-05,
338
+ "loss": 0.0585,
339
+ "step": 470
340
+ },
341
+ {
342
+ "epoch": 0.1911589008363202,
343
+ "grad_norm": 1.1783093214035034,
344
+ "learning_rate": 9.58e-05,
345
+ "loss": 0.0509,
346
+ "step": 480
347
+ },
348
+ {
349
+ "epoch": 0.19514137793707687,
350
+ "grad_norm": 0.9017106294631958,
351
+ "learning_rate": 9.78e-05,
352
+ "loss": 0.0558,
353
+ "step": 490
354
+ },
355
+ {
356
+ "epoch": 0.19912385503783353,
357
+ "grad_norm": 1.0761146545410156,
358
+ "learning_rate": 9.98e-05,
359
+ "loss": 0.0519,
360
+ "step": 500
361
+ },
362
+ {
363
+ "epoch": 0.2031063321385902,
364
+ "grad_norm": 0.5883626937866211,
365
+ "learning_rate": 9.9999778549206e-05,
366
+ "loss": 0.0527,
367
+ "step": 510
368
+ },
369
+ {
370
+ "epoch": 0.20708880923934686,
371
+ "grad_norm": 0.6159175634384155,
372
+ "learning_rate": 9.999901304280685e-05,
373
+ "loss": 0.0488,
374
+ "step": 520
375
+ },
376
+ {
377
+ "epoch": 0.21107128634010355,
378
+ "grad_norm": 0.6942933797836304,
379
+ "learning_rate": 9.999770075521164e-05,
380
+ "loss": 0.0524,
381
+ "step": 530
382
+ },
383
+ {
384
+ "epoch": 0.21505376344086022,
385
+ "grad_norm": 1.0295978784561157,
386
+ "learning_rate": 9.99958417007713e-05,
387
+ "loss": 0.049,
388
+ "step": 540
389
+ },
390
+ {
391
+ "epoch": 0.21903624054161688,
392
+ "grad_norm": 0.7261350154876709,
393
+ "learning_rate": 9.999343589981615e-05,
394
+ "loss": 0.0467,
395
+ "step": 550
396
+ },
397
+ {
398
+ "epoch": 0.22301871764237355,
399
+ "grad_norm": 0.7238365411758423,
400
+ "learning_rate": 9.999048337865568e-05,
401
+ "loss": 0.0477,
402
+ "step": 560
403
+ },
404
+ {
405
+ "epoch": 0.2270011947431302,
406
+ "grad_norm": 1.1104141473770142,
407
+ "learning_rate": 9.998698416957815e-05,
408
+ "loss": 0.0519,
409
+ "step": 570
410
+ },
411
+ {
412
+ "epoch": 0.2309836718438869,
413
+ "grad_norm": 1.2086730003356934,
414
+ "learning_rate": 9.998293831085037e-05,
415
+ "loss": 0.0524,
416
+ "step": 580
417
+ },
418
+ {
419
+ "epoch": 0.23496614894464357,
420
+ "grad_norm": 0.5849373936653137,
421
+ "learning_rate": 9.997834584671719e-05,
422
+ "loss": 0.0514,
423
+ "step": 590
424
+ },
425
+ {
426
+ "epoch": 0.23894862604540024,
427
+ "grad_norm": 0.9337085485458374,
428
+ "learning_rate": 9.997320682740107e-05,
429
+ "loss": 0.0565,
430
+ "step": 600
431
+ },
432
+ {
433
+ "epoch": 0.2429311031461569,
434
+ "grad_norm": 0.7592830061912537,
435
+ "learning_rate": 9.996752130910149e-05,
436
+ "loss": 0.0521,
437
+ "step": 610
438
+ },
439
+ {
440
+ "epoch": 0.24691358024691357,
441
+ "grad_norm": 0.8495620489120483,
442
+ "learning_rate": 9.99612893539944e-05,
443
+ "loss": 0.0501,
444
+ "step": 620
445
+ },
446
+ {
447
+ "epoch": 0.25089605734767023,
448
+ "grad_norm": 0.44615253806114197,
449
+ "learning_rate": 9.995451103023144e-05,
450
+ "loss": 0.0428,
451
+ "step": 630
452
+ },
453
+ {
454
+ "epoch": 0.2548785344484269,
455
+ "grad_norm": 1.046374797821045,
456
+ "learning_rate": 9.994718641193928e-05,
457
+ "loss": 0.0459,
458
+ "step": 640
459
+ },
460
+ {
461
+ "epoch": 0.2588610115491836,
462
+ "grad_norm": 0.7210455536842346,
463
+ "learning_rate": 9.993931557921874e-05,
464
+ "loss": 0.0508,
465
+ "step": 650
466
+ },
467
+ {
468
+ "epoch": 0.2628434886499403,
469
+ "grad_norm": 0.7506885528564453,
470
+ "learning_rate": 9.993089861814402e-05,
471
+ "loss": 0.0499,
472
+ "step": 660
473
+ },
474
+ {
475
+ "epoch": 0.26682596575069695,
476
+ "grad_norm": 0.6320833563804626,
477
+ "learning_rate": 9.992193562076166e-05,
478
+ "loss": 0.0579,
479
+ "step": 670
480
+ },
481
+ {
482
+ "epoch": 0.2708084428514536,
483
+ "grad_norm": 0.48116907477378845,
484
+ "learning_rate": 9.991242668508954e-05,
485
+ "loss": 0.0533,
486
+ "step": 680
487
+ },
488
+ {
489
+ "epoch": 0.2747909199522103,
490
+ "grad_norm": 0.5594539046287537,
491
+ "learning_rate": 9.990237191511587e-05,
492
+ "loss": 0.0479,
493
+ "step": 690
494
+ },
495
+ {
496
+ "epoch": 0.27877339705296694,
497
+ "grad_norm": 0.6126905083656311,
498
+ "learning_rate": 9.989177142079802e-05,
499
+ "loss": 0.0449,
500
+ "step": 700
501
+ },
502
+ {
503
+ "epoch": 0.2827558741537236,
504
+ "grad_norm": 0.8031131625175476,
505
+ "learning_rate": 9.988062531806126e-05,
506
+ "loss": 0.0485,
507
+ "step": 710
508
+ },
509
+ {
510
+ "epoch": 0.2867383512544803,
511
+ "grad_norm": 0.4269757568836212,
512
+ "learning_rate": 9.986893372879762e-05,
513
+ "loss": 0.0463,
514
+ "step": 720
515
+ },
516
+ {
517
+ "epoch": 0.29072082835523694,
518
+ "grad_norm": 0.8300117254257202,
519
+ "learning_rate": 9.985669678086443e-05,
520
+ "loss": 0.0505,
521
+ "step": 730
522
+ },
523
+ {
524
+ "epoch": 0.2947033054559936,
525
+ "grad_norm": 0.6425809264183044,
526
+ "learning_rate": 9.984391460808298e-05,
527
+ "loss": 0.0407,
528
+ "step": 740
529
+ },
530
+ {
531
+ "epoch": 0.2986857825567503,
532
+ "grad_norm": 0.4019189774990082,
533
+ "learning_rate": 9.983058735023709e-05,
534
+ "loss": 0.044,
535
+ "step": 750
536
+ },
537
+ {
538
+ "epoch": 0.302668259657507,
539
+ "grad_norm": 1.2118290662765503,
540
+ "learning_rate": 9.98167151530715e-05,
541
+ "loss": 0.0435,
542
+ "step": 760
543
+ },
544
+ {
545
+ "epoch": 0.30665073675826365,
546
+ "grad_norm": 0.4535205662250519,
547
+ "learning_rate": 9.980229816829034e-05,
548
+ "loss": 0.043,
549
+ "step": 770
550
+ },
551
+ {
552
+ "epoch": 0.3106332138590203,
553
+ "grad_norm": 0.7236475348472595,
554
+ "learning_rate": 9.978733655355544e-05,
555
+ "loss": 0.0479,
556
+ "step": 780
557
+ },
558
+ {
559
+ "epoch": 0.314615690959777,
560
+ "grad_norm": 1.045904278755188,
561
+ "learning_rate": 9.977183047248464e-05,
562
+ "loss": 0.0458,
563
+ "step": 790
564
+ },
565
+ {
566
+ "epoch": 0.31859816806053365,
567
+ "grad_norm": 0.8453964591026306,
568
+ "learning_rate": 9.975578009464992e-05,
569
+ "loss": 0.0489,
570
+ "step": 800
571
+ },
572
+ {
573
+ "epoch": 0.3225806451612903,
574
+ "grad_norm": 0.8722575902938843,
575
+ "learning_rate": 9.97391855955757e-05,
576
+ "loss": 0.0496,
577
+ "step": 810
578
+ },
579
+ {
580
+ "epoch": 0.326563122262047,
581
+ "grad_norm": 0.5823293328285217,
582
+ "learning_rate": 9.972204715673669e-05,
583
+ "loss": 0.0444,
584
+ "step": 820
585
+ },
586
+ {
587
+ "epoch": 0.33054559936280364,
588
+ "grad_norm": 0.426830917596817,
589
+ "learning_rate": 9.970436496555617e-05,
590
+ "loss": 0.0429,
591
+ "step": 830
592
+ },
593
+ {
594
+ "epoch": 0.3345280764635603,
595
+ "grad_norm": 0.6528297066688538,
596
+ "learning_rate": 9.968613921540373e-05,
597
+ "loss": 0.0462,
598
+ "step": 840
599
+ },
600
+ {
601
+ "epoch": 0.33851055356431703,
602
+ "grad_norm": 0.8501228094100952,
603
+ "learning_rate": 9.966737010559326e-05,
604
+ "loss": 0.0483,
605
+ "step": 850
606
+ },
607
+ {
608
+ "epoch": 0.3424930306650737,
609
+ "grad_norm": 0.6673285365104675,
610
+ "learning_rate": 9.964805784138072e-05,
611
+ "loss": 0.0483,
612
+ "step": 860
613
+ },
614
+ {
615
+ "epoch": 0.34647550776583036,
616
+ "grad_norm": 0.6113673448562622,
617
+ "learning_rate": 9.962820263396195e-05,
618
+ "loss": 0.0439,
619
+ "step": 870
620
+ },
621
+ {
622
+ "epoch": 0.350457984866587,
623
+ "grad_norm": 0.8236457705497742,
624
+ "learning_rate": 9.960780470047033e-05,
625
+ "loss": 0.0425,
626
+ "step": 880
627
+ },
628
+ {
629
+ "epoch": 0.3544404619673437,
630
+ "grad_norm": 0.4798528552055359,
631
+ "learning_rate": 9.958686426397437e-05,
632
+ "loss": 0.0397,
633
+ "step": 890
634
+ },
635
+ {
636
+ "epoch": 0.35842293906810035,
637
+ "grad_norm": 0.5209172964096069,
638
+ "learning_rate": 9.956538155347534e-05,
639
+ "loss": 0.0496,
640
+ "step": 900
641
+ },
642
+ {
643
+ "epoch": 0.362405416168857,
644
+ "grad_norm": 0.549393892288208,
645
+ "learning_rate": 9.95433568039047e-05,
646
+ "loss": 0.0388,
647
+ "step": 910
648
+ },
649
+ {
650
+ "epoch": 0.3663878932696137,
651
+ "grad_norm": 0.6078306436538696,
652
+ "learning_rate": 9.952079025612162e-05,
653
+ "loss": 0.0467,
654
+ "step": 920
655
+ },
656
+ {
657
+ "epoch": 0.37037037037037035,
658
+ "grad_norm": 0.47643283009529114,
659
+ "learning_rate": 9.949768215691022e-05,
660
+ "loss": 0.0392,
661
+ "step": 930
662
+ },
663
+ {
664
+ "epoch": 0.374352847471127,
665
+ "grad_norm": 0.5230958461761475,
666
+ "learning_rate": 9.9474032758977e-05,
667
+ "loss": 0.0414,
668
+ "step": 940
669
+ },
670
+ {
671
+ "epoch": 0.37833532457188374,
672
+ "grad_norm": 0.5619562864303589,
673
+ "learning_rate": 9.944984232094794e-05,
674
+ "loss": 0.0448,
675
+ "step": 950
676
+ },
677
+ {
678
+ "epoch": 0.3823178016726404,
679
+ "grad_norm": 0.7934162616729736,
680
+ "learning_rate": 9.942511110736584e-05,
681
+ "loss": 0.0389,
682
+ "step": 960
683
+ },
684
+ {
685
+ "epoch": 0.38630027877339707,
686
+ "grad_norm": 0.7610427737236023,
687
+ "learning_rate": 9.939983938868726e-05,
688
+ "loss": 0.0383,
689
+ "step": 970
690
+ },
691
+ {
692
+ "epoch": 0.39028275587415373,
693
+ "grad_norm": 0.6311072111129761,
694
+ "learning_rate": 9.93740274412797e-05,
695
+ "loss": 0.0449,
696
+ "step": 980
697
+ },
698
+ {
699
+ "epoch": 0.3942652329749104,
700
+ "grad_norm": 0.3925153613090515,
701
+ "learning_rate": 9.934767554741846e-05,
702
+ "loss": 0.043,
703
+ "step": 990
704
+ },
705
+ {
706
+ "epoch": 0.39824771007566706,
707
+ "grad_norm": 0.4677866995334625,
708
+ "learning_rate": 9.932078399528361e-05,
709
+ "loss": 0.0414,
710
+ "step": 1000
711
+ },
712
+ {
713
+ "epoch": 0.4022301871764237,
714
+ "grad_norm": 0.38546767830848694,
715
+ "learning_rate": 9.929335307895689e-05,
716
+ "loss": 0.0342,
717
+ "step": 1010
718
+ },
719
+ {
720
+ "epoch": 0.4062126642771804,
721
+ "grad_norm": 0.5453174710273743,
722
+ "learning_rate": 9.926538309841839e-05,
723
+ "loss": 0.0421,
724
+ "step": 1020
725
+ },
726
+ {
727
+ "epoch": 0.41019514137793706,
728
+ "grad_norm": 0.34743279218673706,
729
+ "learning_rate": 9.923687435954334e-05,
730
+ "loss": 0.0387,
731
+ "step": 1030
732
+ },
733
+ {
734
+ "epoch": 0.4141776184786937,
735
+ "grad_norm": 0.554025411605835,
736
+ "learning_rate": 9.920782717409873e-05,
737
+ "loss": 0.0469,
738
+ "step": 1040
739
+ },
740
+ {
741
+ "epoch": 0.41816009557945044,
742
+ "grad_norm": 0.6354169249534607,
743
+ "learning_rate": 9.917824185973994e-05,
744
+ "loss": 0.0411,
745
+ "step": 1050
746
+ },
747
+ {
748
+ "epoch": 0.4221425726802071,
749
+ "grad_norm": 0.46389538049697876,
750
+ "learning_rate": 9.914811874000723e-05,
751
+ "loss": 0.0447,
752
+ "step": 1060
753
+ },
754
+ {
755
+ "epoch": 0.42612504978096377,
756
+ "grad_norm": 0.8935972452163696,
757
+ "learning_rate": 9.911745814432218e-05,
758
+ "loss": 0.0454,
759
+ "step": 1070
760
+ },
761
+ {
762
+ "epoch": 0.43010752688172044,
763
+ "grad_norm": 1.0306254625320435,
764
+ "learning_rate": 9.90862604079842e-05,
765
+ "loss": 0.0438,
766
+ "step": 1080
767
+ },
768
+ {
769
+ "epoch": 0.4340900039824771,
770
+ "grad_norm": 0.5918930768966675,
771
+ "learning_rate": 9.90545258721667e-05,
772
+ "loss": 0.0389,
773
+ "step": 1090
774
+ },
775
+ {
776
+ "epoch": 0.43807248108323377,
777
+ "grad_norm": 0.5646054148674011,
778
+ "learning_rate": 9.90222548839135e-05,
779
+ "loss": 0.0395,
780
+ "step": 1100
781
+ },
782
+ {
783
+ "epoch": 0.44205495818399043,
784
+ "grad_norm": 0.5279185175895691,
785
+ "learning_rate": 9.898944779613495e-05,
786
+ "loss": 0.0438,
787
+ "step": 1110
788
+ },
789
+ {
790
+ "epoch": 0.4460374352847471,
791
+ "grad_norm": 0.41257861256599426,
792
+ "learning_rate": 9.89561049676041e-05,
793
+ "loss": 0.0397,
794
+ "step": 1120
795
+ },
796
+ {
797
+ "epoch": 0.45001991238550376,
798
+ "grad_norm": 0.36252203583717346,
799
+ "learning_rate": 9.89222267629528e-05,
800
+ "loss": 0.043,
801
+ "step": 1130
802
+ },
803
+ {
804
+ "epoch": 0.4540023894862604,
805
+ "grad_norm": 0.8424637913703918,
806
+ "learning_rate": 9.888781355266763e-05,
807
+ "loss": 0.0433,
808
+ "step": 1140
809
+ },
810
+ {
811
+ "epoch": 0.45798486658701715,
812
+ "grad_norm": 0.5604090690612793,
813
+ "learning_rate": 9.885286571308598e-05,
814
+ "loss": 0.0452,
815
+ "step": 1150
816
+ },
817
+ {
818
+ "epoch": 0.4619673436877738,
819
+ "grad_norm": 0.7178535461425781,
820
+ "learning_rate": 9.881738362639182e-05,
821
+ "loss": 0.0401,
822
+ "step": 1160
823
+ },
824
+ {
825
+ "epoch": 0.4659498207885305,
826
+ "grad_norm": 0.7291957139968872,
827
+ "learning_rate": 9.878136768061154e-05,
828
+ "loss": 0.0368,
829
+ "step": 1170
830
+ },
831
+ {
832
+ "epoch": 0.46993229788928714,
833
+ "grad_norm": 0.48816534876823425,
834
+ "learning_rate": 9.874481826960979e-05,
835
+ "loss": 0.039,
836
+ "step": 1180
837
+ },
838
+ {
839
+ "epoch": 0.4739147749900438,
840
+ "grad_norm": 0.813832700252533,
841
+ "learning_rate": 9.870773579308503e-05,
842
+ "loss": 0.0417,
843
+ "step": 1190
844
+ },
845
+ {
846
+ "epoch": 0.4778972520908005,
847
+ "grad_norm": 0.5479532480239868,
848
+ "learning_rate": 9.867012065656533e-05,
849
+ "loss": 0.0428,
850
+ "step": 1200
851
+ },
852
+ {
853
+ "epoch": 0.48187972919155714,
854
+ "grad_norm": 0.7760400772094727,
855
+ "learning_rate": 9.863197327140376e-05,
856
+ "loss": 0.0454,
857
+ "step": 1210
858
+ },
859
+ {
860
+ "epoch": 0.4858622062923138,
861
+ "grad_norm": 0.6200589537620544,
862
+ "learning_rate": 9.859329405477403e-05,
863
+ "loss": 0.0377,
864
+ "step": 1220
865
+ },
866
+ {
867
+ "epoch": 0.48984468339307047,
868
+ "grad_norm": 0.7709351778030396,
869
+ "learning_rate": 9.855408342966585e-05,
870
+ "loss": 0.045,
871
+ "step": 1230
872
+ },
873
+ {
874
+ "epoch": 0.49382716049382713,
875
+ "grad_norm": 0.44105854630470276,
876
+ "learning_rate": 9.851434182488033e-05,
877
+ "loss": 0.0374,
878
+ "step": 1240
879
+ },
880
+ {
881
+ "epoch": 0.49780963759458385,
882
+ "grad_norm": 0.4224795401096344,
883
+ "learning_rate": 9.84740696750253e-05,
884
+ "loss": 0.0403,
885
+ "step": 1250
886
+ },
887
+ {
888
+ "epoch": 0.5017921146953405,
889
+ "grad_norm": 0.49829116463661194,
890
+ "learning_rate": 9.843326742051055e-05,
891
+ "loss": 0.0427,
892
+ "step": 1260
893
+ },
894
+ {
895
+ "epoch": 0.5057745917960972,
896
+ "grad_norm": 0.42388513684272766,
897
+ "learning_rate": 9.839193550754297e-05,
898
+ "loss": 0.0407,
899
+ "step": 1270
900
+ },
901
+ {
902
+ "epoch": 0.5097570688968538,
903
+ "grad_norm": 0.8282580375671387,
904
+ "learning_rate": 9.835007438812177e-05,
905
+ "loss": 0.0422,
906
+ "step": 1280
907
+ },
908
+ {
909
+ "epoch": 0.5137395459976105,
910
+ "grad_norm": 0.7779645323753357,
911
+ "learning_rate": 9.830768452003341e-05,
912
+ "loss": 0.044,
913
+ "step": 1290
914
+ },
915
+ {
916
+ "epoch": 0.5177220230983672,
917
+ "grad_norm": 0.6992411017417908,
918
+ "learning_rate": 9.826476636684671e-05,
919
+ "loss": 0.0422,
920
+ "step": 1300
921
+ },
922
+ {
923
+ "epoch": 0.5217045001991238,
924
+ "grad_norm": 0.5714141130447388,
925
+ "learning_rate": 9.822132039790773e-05,
926
+ "loss": 0.0394,
927
+ "step": 1310
928
+ },
929
+ {
930
+ "epoch": 0.5256869772998806,
931
+ "grad_norm": 0.8175871968269348,
932
+ "learning_rate": 9.817734708833461e-05,
933
+ "loss": 0.0498,
934
+ "step": 1320
935
+ },
936
+ {
937
+ "epoch": 0.5296694544006372,
938
+ "grad_norm": 0.6209958791732788,
939
+ "learning_rate": 9.813284691901243e-05,
940
+ "loss": 0.04,
941
+ "step": 1330
942
+ },
943
+ {
944
+ "epoch": 0.5336519315013939,
945
+ "grad_norm": 0.39030686020851135,
946
+ "learning_rate": 9.808782037658792e-05,
947
+ "loss": 0.0414,
948
+ "step": 1340
949
+ },
950
+ {
951
+ "epoch": 0.5376344086021505,
952
+ "grad_norm": 0.5858843922615051,
953
+ "learning_rate": 9.804226795346411e-05,
954
+ "loss": 0.0384,
955
+ "step": 1350
956
+ },
957
+ {
958
+ "epoch": 0.5416168857029072,
959
+ "grad_norm": 0.812506914138794,
960
+ "learning_rate": 9.799619014779503e-05,
961
+ "loss": 0.0428,
962
+ "step": 1360
963
+ },
964
+ {
965
+ "epoch": 0.5455993628036638,
966
+ "grad_norm": 0.5438332557678223,
967
+ "learning_rate": 9.794958746348013e-05,
968
+ "loss": 0.0344,
969
+ "step": 1370
970
+ },
971
+ {
972
+ "epoch": 0.5495818399044206,
973
+ "grad_norm": 0.4357604384422302,
974
+ "learning_rate": 9.790246041015896e-05,
975
+ "loss": 0.0429,
976
+ "step": 1380
977
+ },
978
+ {
979
+ "epoch": 0.5535643170051773,
980
+ "grad_norm": 0.6829274892807007,
981
+ "learning_rate": 9.785480950320538e-05,
982
+ "loss": 0.0352,
983
+ "step": 1390
984
+ },
985
+ {
986
+ "epoch": 0.5575467941059339,
987
+ "grad_norm": 0.46741291880607605,
988
+ "learning_rate": 9.78066352637221e-05,
989
+ "loss": 0.0387,
990
+ "step": 1400
991
+ },
992
+ {
993
+ "epoch": 0.5615292712066906,
994
+ "grad_norm": 0.8722621202468872,
995
+ "learning_rate": 9.775793821853488e-05,
996
+ "loss": 0.0364,
997
+ "step": 1410
998
+ },
999
+ {
1000
+ "epoch": 0.5655117483074472,
1001
+ "grad_norm": 0.6563253402709961,
1002
+ "learning_rate": 9.77087189001868e-05,
1003
+ "loss": 0.0409,
1004
+ "step": 1420
1005
+ },
1006
+ {
1007
+ "epoch": 0.5694942254082039,
1008
+ "grad_norm": 0.6959318518638611,
1009
+ "learning_rate": 9.765897784693243e-05,
1010
+ "loss": 0.0349,
1011
+ "step": 1430
1012
+ },
1013
+ {
1014
+ "epoch": 0.5734767025089605,
1015
+ "grad_norm": 0.34205693006515503,
1016
+ "learning_rate": 9.760871560273197e-05,
1017
+ "loss": 0.036,
1018
+ "step": 1440
1019
+ },
1020
+ {
1021
+ "epoch": 0.5774591796097173,
1022
+ "grad_norm": 0.7076167464256287,
1023
+ "learning_rate": 9.755793271724526e-05,
1024
+ "loss": 0.0413,
1025
+ "step": 1450
1026
+ },
1027
+ {
1028
+ "epoch": 0.5814416567104739,
1029
+ "grad_norm": 0.6335232257843018,
1030
+ "learning_rate": 9.750662974582584e-05,
1031
+ "loss": 0.0392,
1032
+ "step": 1460
1033
+ },
1034
+ {
1035
+ "epoch": 0.5854241338112306,
1036
+ "grad_norm": 0.7028225660324097,
1037
+ "learning_rate": 9.745480724951473e-05,
1038
+ "loss": 0.0391,
1039
+ "step": 1470
1040
+ },
1041
+ {
1042
+ "epoch": 0.5894066109119872,
1043
+ "grad_norm": 0.5494594573974609,
1044
+ "learning_rate": 9.740246579503447e-05,
1045
+ "loss": 0.0469,
1046
+ "step": 1480
1047
+ },
1048
+ {
1049
+ "epoch": 0.5933890880127439,
1050
+ "grad_norm": 0.3793744146823883,
1051
+ "learning_rate": 9.734960595478284e-05,
1052
+ "loss": 0.0372,
1053
+ "step": 1490
1054
+ },
1055
+ {
1056
+ "epoch": 0.5973715651135006,
1057
+ "grad_norm": 0.6068878769874573,
1058
+ "learning_rate": 9.729622830682657e-05,
1059
+ "loss": 0.0443,
1060
+ "step": 1500
1061
+ },
1062
+ {
1063
+ "epoch": 0.6013540422142573,
1064
+ "grad_norm": 0.6348056197166443,
1065
+ "learning_rate": 9.724233343489504e-05,
1066
+ "loss": 0.0388,
1067
+ "step": 1510
1068
+ },
1069
+ {
1070
+ "epoch": 0.605336519315014,
1071
+ "grad_norm": 0.7185152173042297,
1072
+ "learning_rate": 9.718792192837396e-05,
1073
+ "loss": 0.0394,
1074
+ "step": 1520
1075
+ },
1076
+ {
1077
+ "epoch": 0.6093189964157706,
1078
+ "grad_norm": 0.29940369725227356,
1079
+ "learning_rate": 9.713299438229886e-05,
1080
+ "loss": 0.0373,
1081
+ "step": 1530
1082
+ },
1083
+ {
1084
+ "epoch": 0.6133014735165273,
1085
+ "grad_norm": 0.608589768409729,
1086
+ "learning_rate": 9.707755139734855e-05,
1087
+ "loss": 0.0444,
1088
+ "step": 1540
1089
+ },
1090
+ {
1091
+ "epoch": 0.6172839506172839,
1092
+ "grad_norm": 0.6535037755966187,
1093
+ "learning_rate": 9.702159357983866e-05,
1094
+ "loss": 0.0418,
1095
+ "step": 1550
1096
+ },
1097
+ {
1098
+ "epoch": 0.6212664277180406,
1099
+ "grad_norm": 0.389615923166275,
1100
+ "learning_rate": 9.696512154171492e-05,
1101
+ "loss": 0.0394,
1102
+ "step": 1560
1103
+ },
1104
+ {
1105
+ "epoch": 0.6252489048187972,
1106
+ "grad_norm": 0.5415939092636108,
1107
+ "learning_rate": 9.690813590054645e-05,
1108
+ "loss": 0.0374,
1109
+ "step": 1570
1110
+ },
1111
+ {
1112
+ "epoch": 0.629231381919554,
1113
+ "grad_norm": 0.6598183512687683,
1114
+ "learning_rate": 9.685063727951914e-05,
1115
+ "loss": 0.0364,
1116
+ "step": 1580
1117
+ },
1118
+ {
1119
+ "epoch": 0.6332138590203107,
1120
+ "grad_norm": 0.4346562325954437,
1121
+ "learning_rate": 9.679262630742865e-05,
1122
+ "loss": 0.0368,
1123
+ "step": 1590
1124
+ },
1125
+ {
1126
+ "epoch": 0.6371963361210673,
1127
+ "grad_norm": 0.7525681257247925,
1128
+ "learning_rate": 9.673410361867373e-05,
1129
+ "loss": 0.037,
1130
+ "step": 1600
1131
+ },
1132
+ {
1133
+ "epoch": 0.641178813221824,
1134
+ "grad_norm": 0.4621691405773163,
1135
+ "learning_rate": 9.667506985324909e-05,
1136
+ "loss": 0.0313,
1137
+ "step": 1610
1138
+ },
1139
+ {
1140
+ "epoch": 0.6451612903225806,
1141
+ "grad_norm": 0.5039979219436646,
1142
+ "learning_rate": 9.661552565673855e-05,
1143
+ "loss": 0.0309,
1144
+ "step": 1620
1145
+ },
1146
+ {
1147
+ "epoch": 0.6491437674233373,
1148
+ "grad_norm": 0.29736581444740295,
1149
+ "learning_rate": 9.655547168030789e-05,
1150
+ "loss": 0.0369,
1151
+ "step": 1630
1152
+ },
1153
+ {
1154
+ "epoch": 0.653126244524094,
1155
+ "grad_norm": 0.4153502583503723,
1156
+ "learning_rate": 9.649490858069777e-05,
1157
+ "loss": 0.0349,
1158
+ "step": 1640
1159
+ },
1160
+ {
1161
+ "epoch": 0.6571087216248507,
1162
+ "grad_norm": 0.64329993724823,
1163
+ "learning_rate": 9.643383702021658e-05,
1164
+ "loss": 0.0422,
1165
+ "step": 1650
1166
+ },
1167
+ {
1168
+ "epoch": 0.6610911987256073,
1169
+ "grad_norm": 0.39090660214424133,
1170
+ "learning_rate": 9.637225766673307e-05,
1171
+ "loss": 0.0332,
1172
+ "step": 1660
1173
+ },
1174
+ {
1175
+ "epoch": 0.665073675826364,
1176
+ "grad_norm": 0.4412473440170288,
1177
+ "learning_rate": 9.631017119366922e-05,
1178
+ "loss": 0.0374,
1179
+ "step": 1670
1180
+ },
1181
+ {
1182
+ "epoch": 0.6690561529271206,
1183
+ "grad_norm": 0.28402477502822876,
1184
+ "learning_rate": 9.624757827999273e-05,
1185
+ "loss": 0.0364,
1186
+ "step": 1680
1187
+ },
1188
+ {
1189
+ "epoch": 0.6730386300278773,
1190
+ "grad_norm": 0.5480361580848694,
1191
+ "learning_rate": 9.618447961020971e-05,
1192
+ "loss": 0.0381,
1193
+ "step": 1690
1194
+ },
1195
+ {
1196
+ "epoch": 0.6770211071286341,
1197
+ "grad_norm": 0.6572909355163574,
1198
+ "learning_rate": 9.612087587435707e-05,
1199
+ "loss": 0.037,
1200
+ "step": 1700
1201
+ },
1202
+ {
1203
+ "epoch": 0.6810035842293907,
1204
+ "grad_norm": 0.6667490601539612,
1205
+ "learning_rate": 9.605676776799508e-05,
1206
+ "loss": 0.04,
1207
+ "step": 1710
1208
+ },
1209
+ {
1210
+ "epoch": 0.6849860613301474,
1211
+ "grad_norm": 0.29837551712989807,
1212
+ "learning_rate": 9.599215599219973e-05,
1213
+ "loss": 0.0351,
1214
+ "step": 1720
1215
+ },
1216
+ {
1217
+ "epoch": 0.688968538430904,
1218
+ "grad_norm": 0.6289225220680237,
1219
+ "learning_rate": 9.592704125355505e-05,
1220
+ "loss": 0.0377,
1221
+ "step": 1730
1222
+ },
1223
+ {
1224
+ "epoch": 0.6929510155316607,
1225
+ "grad_norm": 0.4489145576953888,
1226
+ "learning_rate": 9.586142426414538e-05,
1227
+ "loss": 0.0391,
1228
+ "step": 1740
1229
+ },
1230
+ {
1231
+ "epoch": 0.6969334926324173,
1232
+ "grad_norm": 0.5505160093307495,
1233
+ "learning_rate": 9.57953057415476e-05,
1234
+ "loss": 0.0371,
1235
+ "step": 1750
1236
+ },
1237
+ {
1238
+ "epoch": 0.700915969733174,
1239
+ "grad_norm": 0.48338383436203003,
1240
+ "learning_rate": 9.572868640882328e-05,
1241
+ "loss": 0.0399,
1242
+ "step": 1760
1243
+ },
1244
+ {
1245
+ "epoch": 0.7048984468339307,
1246
+ "grad_norm": 0.3950949013233185,
1247
+ "learning_rate": 9.56615669945108e-05,
1248
+ "loss": 0.0411,
1249
+ "step": 1770
1250
+ },
1251
+ {
1252
+ "epoch": 0.7088809239346874,
1253
+ "grad_norm": 0.7186228036880493,
1254
+ "learning_rate": 9.55939482326173e-05,
1255
+ "loss": 0.0358,
1256
+ "step": 1780
1257
+ },
1258
+ {
1259
+ "epoch": 0.7128634010354441,
1260
+ "grad_norm": 0.36578911542892456,
1261
+ "learning_rate": 9.552583086261069e-05,
1262
+ "loss": 0.0402,
1263
+ "step": 1790
1264
+ },
1265
+ {
1266
+ "epoch": 0.7168458781362007,
1267
+ "grad_norm": 0.48990169167518616,
1268
+ "learning_rate": 9.545721562941168e-05,
1269
+ "loss": 0.0353,
1270
+ "step": 1800
1271
+ },
1272
+ {
1273
+ "epoch": 0.7208283552369574,
1274
+ "grad_norm": 0.37718433141708374,
1275
+ "learning_rate": 9.538810328338543e-05,
1276
+ "loss": 0.0385,
1277
+ "step": 1810
1278
+ },
1279
+ {
1280
+ "epoch": 0.724810832337714,
1281
+ "grad_norm": 0.40954500436782837,
1282
+ "learning_rate": 9.531849458033349e-05,
1283
+ "loss": 0.0334,
1284
+ "step": 1820
1285
+ },
1286
+ {
1287
+ "epoch": 0.7287933094384708,
1288
+ "grad_norm": 0.6403422951698303,
1289
+ "learning_rate": 9.524839028148547e-05,
1290
+ "loss": 0.0385,
1291
+ "step": 1830
1292
+ },
1293
+ {
1294
+ "epoch": 0.7327757865392274,
1295
+ "grad_norm": 0.4989095628261566,
1296
+ "learning_rate": 9.517779115349077e-05,
1297
+ "loss": 0.0374,
1298
+ "step": 1840
1299
+ },
1300
+ {
1301
+ "epoch": 0.7367582636399841,
1302
+ "grad_norm": 0.4369076192378998,
1303
+ "learning_rate": 9.510669796841014e-05,
1304
+ "loss": 0.0353,
1305
+ "step": 1850
1306
+ },
1307
+ {
1308
+ "epoch": 0.7407407407407407,
1309
+ "grad_norm": 0.6268195509910583,
1310
+ "learning_rate": 9.503511150370727e-05,
1311
+ "loss": 0.0407,
1312
+ "step": 1860
1313
+ },
1314
+ {
1315
+ "epoch": 0.7447232178414974,
1316
+ "grad_norm": 0.40719932317733765,
1317
+ "learning_rate": 9.496303254224024e-05,
1318
+ "loss": 0.037,
1319
+ "step": 1870
1320
+ },
1321
+ {
1322
+ "epoch": 0.748705694942254,
1323
+ "grad_norm": 0.41097119450569153,
1324
+ "learning_rate": 9.489046187225306e-05,
1325
+ "loss": 0.0338,
1326
+ "step": 1880
1327
+ },
1328
+ {
1329
+ "epoch": 0.7526881720430108,
1330
+ "grad_norm": 0.3966093361377716,
1331
+ "learning_rate": 9.481740028736692e-05,
1332
+ "loss": 0.0318,
1333
+ "step": 1890
1334
+ },
1335
+ {
1336
+ "epoch": 0.7566706491437675,
1337
+ "grad_norm": 0.6017296314239502,
1338
+ "learning_rate": 9.474384858657164e-05,
1339
+ "loss": 0.0375,
1340
+ "step": 1900
1341
+ },
1342
+ {
1343
+ "epoch": 0.7606531262445241,
1344
+ "grad_norm": 0.6876013278961182,
1345
+ "learning_rate": 9.466980757421679e-05,
1346
+ "loss": 0.0381,
1347
+ "step": 1910
1348
+ },
1349
+ {
1350
+ "epoch": 0.7646356033452808,
1351
+ "grad_norm": 0.4606328010559082,
1352
+ "learning_rate": 9.459527806000305e-05,
1353
+ "loss": 0.0379,
1354
+ "step": 1920
1355
+ },
1356
+ {
1357
+ "epoch": 0.7686180804460374,
1358
+ "grad_norm": 0.9279587268829346,
1359
+ "learning_rate": 9.452026085897325e-05,
1360
+ "loss": 0.0351,
1361
+ "step": 1930
1362
+ },
1363
+ {
1364
+ "epoch": 0.7726005575467941,
1365
+ "grad_norm": 0.5607491135597229,
1366
+ "learning_rate": 9.444475679150348e-05,
1367
+ "loss": 0.0397,
1368
+ "step": 1940
1369
+ },
1370
+ {
1371
+ "epoch": 0.7765830346475507,
1372
+ "grad_norm": 0.7617636322975159,
1373
+ "learning_rate": 9.436876668329411e-05,
1374
+ "loss": 0.0357,
1375
+ "step": 1950
1376
+ },
1377
+ {
1378
+ "epoch": 0.7805655117483075,
1379
+ "grad_norm": 0.40550336241722107,
1380
+ "learning_rate": 9.429229136536079e-05,
1381
+ "loss": 0.0326,
1382
+ "step": 1960
1383
+ },
1384
+ {
1385
+ "epoch": 0.7845479888490641,
1386
+ "grad_norm": 0.7706924676895142,
1387
+ "learning_rate": 9.421533167402534e-05,
1388
+ "loss": 0.0395,
1389
+ "step": 1970
1390
+ },
1391
+ {
1392
+ "epoch": 0.7885304659498208,
1393
+ "grad_norm": 0.5256568789482117,
1394
+ "learning_rate": 9.413788845090666e-05,
1395
+ "loss": 0.04,
1396
+ "step": 1980
1397
+ },
1398
+ {
1399
+ "epoch": 0.7925129430505775,
1400
+ "grad_norm": 0.4805433452129364,
1401
+ "learning_rate": 9.405996254291136e-05,
1402
+ "loss": 0.0358,
1403
+ "step": 1990
1404
+ },
1405
+ {
1406
+ "epoch": 0.7964954201513341,
1407
+ "grad_norm": 0.4774327874183655,
1408
+ "learning_rate": 9.398155480222474e-05,
1409
+ "loss": 0.0323,
1410
+ "step": 2000
1411
+ },
1412
+ {
1413
+ "epoch": 0.8004778972520908,
1414
+ "grad_norm": 0.42436131834983826,
1415
+ "learning_rate": 9.390266608630128e-05,
1416
+ "loss": 0.0351,
1417
+ "step": 2010
1418
+ },
1419
+ {
1420
+ "epoch": 0.8044603743528475,
1421
+ "grad_norm": 0.43557778000831604,
1422
+ "learning_rate": 9.38232972578553e-05,
1423
+ "loss": 0.0385,
1424
+ "step": 2020
1425
+ },
1426
+ {
1427
+ "epoch": 0.8084428514536042,
1428
+ "grad_norm": 0.47707608342170715,
1429
+ "learning_rate": 9.374344918485164e-05,
1430
+ "loss": 0.0371,
1431
+ "step": 2030
1432
+ },
1433
+ {
1434
+ "epoch": 0.8124253285543608,
1435
+ "grad_norm": 0.4659009873867035,
1436
+ "learning_rate": 9.366312274049602e-05,
1437
+ "loss": 0.0407,
1438
+ "step": 2040
1439
+ },
1440
+ {
1441
+ "epoch": 0.8164078056551175,
1442
+ "grad_norm": 0.4538560211658478,
1443
+ "learning_rate": 9.358231880322554e-05,
1444
+ "loss": 0.0349,
1445
+ "step": 2050
1446
+ },
1447
+ {
1448
+ "epoch": 0.8203902827558741,
1449
+ "grad_norm": 0.6766425967216492,
1450
+ "learning_rate": 9.350103825669916e-05,
1451
+ "loss": 0.0311,
1452
+ "step": 2060
1453
+ },
1454
+ {
1455
+ "epoch": 0.8243727598566308,
1456
+ "grad_norm": 0.3502272963523865,
1457
+ "learning_rate": 9.341928198978787e-05,
1458
+ "loss": 0.0347,
1459
+ "step": 2070
1460
+ },
1461
+ {
1462
+ "epoch": 0.8283552369573874,
1463
+ "grad_norm": 0.5337536931037903,
1464
+ "learning_rate": 9.333705089656512e-05,
1465
+ "loss": 0.03,
1466
+ "step": 2080
1467
+ },
1468
+ {
1469
+ "epoch": 0.8323377140581442,
1470
+ "grad_norm": 0.6512131690979004,
1471
+ "learning_rate": 9.325434587629698e-05,
1472
+ "loss": 0.0329,
1473
+ "step": 2090
1474
+ },
1475
+ {
1476
+ "epoch": 0.8363201911589009,
1477
+ "grad_norm": 0.5638608336448669,
1478
+ "learning_rate": 9.31711678334323e-05,
1479
+ "loss": 0.0311,
1480
+ "step": 2100
1481
+ },
1482
+ {
1483
+ "epoch": 0.8403026682596575,
1484
+ "grad_norm": 0.3121340572834015,
1485
+ "learning_rate": 9.308751767759282e-05,
1486
+ "loss": 0.0319,
1487
+ "step": 2110
1488
+ },
1489
+ {
1490
+ "epoch": 0.8442851453604142,
1491
+ "grad_norm": 0.6195731163024902,
1492
+ "learning_rate": 9.300339632356325e-05,
1493
+ "loss": 0.0318,
1494
+ "step": 2120
1495
+ },
1496
+ {
1497
+ "epoch": 0.8482676224611708,
1498
+ "grad_norm": 0.6970657110214233,
1499
+ "learning_rate": 9.291880469128124e-05,
1500
+ "loss": 0.0337,
1501
+ "step": 2130
1502
+ },
1503
+ {
1504
+ "epoch": 0.8522500995619275,
1505
+ "grad_norm": 0.5600506663322449,
1506
+ "learning_rate": 9.283374370582732e-05,
1507
+ "loss": 0.0371,
1508
+ "step": 2140
1509
+ },
1510
+ {
1511
+ "epoch": 0.8562325766626842,
1512
+ "grad_norm": 0.47252482175827026,
1513
+ "learning_rate": 9.274821429741482e-05,
1514
+ "loss": 0.0399,
1515
+ "step": 2150
1516
+ },
1517
+ {
1518
+ "epoch": 0.8602150537634409,
1519
+ "grad_norm": 0.511959969997406,
1520
+ "learning_rate": 9.266221740137961e-05,
1521
+ "loss": 0.0336,
1522
+ "step": 2160
1523
+ },
1524
+ {
1525
+ "epoch": 0.8641975308641975,
1526
+ "grad_norm": 0.4836970567703247,
1527
+ "learning_rate": 9.257575395817001e-05,
1528
+ "loss": 0.0297,
1529
+ "step": 2170
1530
+ },
1531
+ {
1532
+ "epoch": 0.8681800079649542,
1533
+ "grad_norm": 0.6154841184616089,
1534
+ "learning_rate": 9.248882491333637e-05,
1535
+ "loss": 0.0369,
1536
+ "step": 2180
1537
+ },
1538
+ {
1539
+ "epoch": 0.8721624850657109,
1540
+ "grad_norm": 0.6111750602722168,
1541
+ "learning_rate": 9.240143121752076e-05,
1542
+ "loss": 0.0325,
1543
+ "step": 2190
1544
+ },
1545
+ {
1546
+ "epoch": 0.8761449621664675,
1547
+ "grad_norm": 0.4122430980205536,
1548
+ "learning_rate": 9.23135738264467e-05,
1549
+ "loss": 0.0325,
1550
+ "step": 2200
1551
+ },
1552
+ {
1553
+ "epoch": 0.8801274392672243,
1554
+ "grad_norm": 0.35100454092025757,
1555
+ "learning_rate": 9.222525370090849e-05,
1556
+ "loss": 0.0313,
1557
+ "step": 2210
1558
+ },
1559
+ {
1560
+ "epoch": 0.8841099163679809,
1561
+ "grad_norm": 0.5073012113571167,
1562
+ "learning_rate": 9.213647180676088e-05,
1563
+ "loss": 0.0337,
1564
+ "step": 2220
1565
+ },
1566
+ {
1567
+ "epoch": 0.8880923934687376,
1568
+ "grad_norm": 0.6666367650032043,
1569
+ "learning_rate": 9.204722911490846e-05,
1570
+ "loss": 0.0357,
1571
+ "step": 2230
1572
+ },
1573
+ {
1574
+ "epoch": 0.8920748705694942,
1575
+ "grad_norm": 0.43829846382141113,
1576
+ "learning_rate": 9.1957526601295e-05,
1577
+ "loss": 0.0309,
1578
+ "step": 2240
1579
+ },
1580
+ {
1581
+ "epoch": 0.8960573476702509,
1582
+ "grad_norm": 0.41505926847457886,
1583
+ "learning_rate": 9.186736524689281e-05,
1584
+ "loss": 0.0362,
1585
+ "step": 2250
1586
+ },
1587
+ {
1588
+ "epoch": 0.9000398247710075,
1589
+ "grad_norm": 0.585198700428009,
1590
+ "learning_rate": 9.177674603769204e-05,
1591
+ "loss": 0.0373,
1592
+ "step": 2260
1593
+ },
1594
+ {
1595
+ "epoch": 0.9040223018717642,
1596
+ "grad_norm": 0.5691582560539246,
1597
+ "learning_rate": 9.168566996468983e-05,
1598
+ "loss": 0.0297,
1599
+ "step": 2270
1600
+ },
1601
+ {
1602
+ "epoch": 0.9080047789725209,
1603
+ "grad_norm": 0.568215012550354,
1604
+ "learning_rate": 9.159413802387951e-05,
1605
+ "loss": 0.0375,
1606
+ "step": 2280
1607
+ },
1608
+ {
1609
+ "epoch": 0.9119872560732776,
1610
+ "grad_norm": 0.41294440627098083,
1611
+ "learning_rate": 9.150215121623974e-05,
1612
+ "loss": 0.0348,
1613
+ "step": 2290
1614
+ },
1615
+ {
1616
+ "epoch": 0.9159697331740343,
1617
+ "grad_norm": 0.5145776271820068,
1618
+ "learning_rate": 9.140971054772349e-05,
1619
+ "loss": 0.0318,
1620
+ "step": 2300
1621
+ },
1622
+ {
1623
+ "epoch": 0.9199522102747909,
1624
+ "grad_norm": 0.4824906289577484,
1625
+ "learning_rate": 9.131681702924713e-05,
1626
+ "loss": 0.0383,
1627
+ "step": 2310
1628
+ },
1629
+ {
1630
+ "epoch": 0.9239346873755476,
1631
+ "grad_norm": 0.5607543587684631,
1632
+ "learning_rate": 9.122347167667926e-05,
1633
+ "loss": 0.0288,
1634
+ "step": 2320
1635
+ },
1636
+ {
1637
+ "epoch": 0.9279171644763042,
1638
+ "grad_norm": 0.7051650285720825,
1639
+ "learning_rate": 9.112967551082973e-05,
1640
+ "loss": 0.0393,
1641
+ "step": 2330
1642
+ },
1643
+ {
1644
+ "epoch": 0.931899641577061,
1645
+ "grad_norm": 0.4049379825592041,
1646
+ "learning_rate": 9.103542955743835e-05,
1647
+ "loss": 0.0322,
1648
+ "step": 2340
1649
+ },
1650
+ {
1651
+ "epoch": 0.9358821186778176,
1652
+ "grad_norm": 0.3805215060710907,
1653
+ "learning_rate": 9.094073484716381e-05,
1654
+ "loss": 0.0357,
1655
+ "step": 2350
1656
+ },
1657
+ {
1658
+ "epoch": 0.9398645957785743,
1659
+ "grad_norm": 0.44789791107177734,
1660
+ "learning_rate": 9.084559241557226e-05,
1661
+ "loss": 0.0291,
1662
+ "step": 2360
1663
+ },
1664
+ {
1665
+ "epoch": 0.9438470728793309,
1666
+ "grad_norm": 0.6356040835380554,
1667
+ "learning_rate": 9.075000330312608e-05,
1668
+ "loss": 0.0322,
1669
+ "step": 2370
1670
+ },
1671
+ {
1672
+ "epoch": 0.9478295499800876,
1673
+ "grad_norm": 0.504687488079071,
1674
+ "learning_rate": 9.065396855517253e-05,
1675
+ "loss": 0.03,
1676
+ "step": 2380
1677
+ },
1678
+ {
1679
+ "epoch": 0.9518120270808443,
1680
+ "grad_norm": 0.4105326533317566,
1681
+ "learning_rate": 9.055748922193219e-05,
1682
+ "loss": 0.0349,
1683
+ "step": 2390
1684
+ },
1685
+ {
1686
+ "epoch": 0.955794504181601,
1687
+ "grad_norm": 0.4912046790122986,
1688
+ "learning_rate": 9.046056635848761e-05,
1689
+ "loss": 0.0384,
1690
+ "step": 2400
1691
+ },
1692
+ {
1693
+ "epoch": 0.9597769812823577,
1694
+ "grad_norm": 0.7549566030502319,
1695
+ "learning_rate": 9.036320102477169e-05,
1696
+ "loss": 0.0327,
1697
+ "step": 2410
1698
+ },
1699
+ {
1700
+ "epoch": 0.9637594583831143,
1701
+ "grad_norm": 0.36022019386291504,
1702
+ "learning_rate": 9.02653942855561e-05,
1703
+ "loss": 0.0331,
1704
+ "step": 2420
1705
+ },
1706
+ {
1707
+ "epoch": 0.967741935483871,
1708
+ "grad_norm": 0.48688042163848877,
1709
+ "learning_rate": 9.016714721043971e-05,
1710
+ "loss": 0.0318,
1711
+ "step": 2430
1712
+ },
1713
+ {
1714
+ "epoch": 0.9717244125846276,
1715
+ "grad_norm": 0.3235294222831726,
1716
+ "learning_rate": 9.006846087383675e-05,
1717
+ "loss": 0.0327,
1718
+ "step": 2440
1719
+ },
1720
+ {
1721
+ "epoch": 0.9757068896853843,
1722
+ "grad_norm": 0.5939168334007263,
1723
+ "learning_rate": 8.996933635496523e-05,
1724
+ "loss": 0.0314,
1725
+ "step": 2450
1726
+ },
1727
+ {
1728
+ "epoch": 0.9796893667861409,
1729
+ "grad_norm": 0.7387416958808899,
1730
+ "learning_rate": 8.986977473783498e-05,
1731
+ "loss": 0.0374,
1732
+ "step": 2460
1733
+ },
1734
+ {
1735
+ "epoch": 0.9836718438868977,
1736
+ "grad_norm": 0.48453521728515625,
1737
+ "learning_rate": 8.97697771112359e-05,
1738
+ "loss": 0.0317,
1739
+ "step": 2470
1740
+ },
1741
+ {
1742
+ "epoch": 0.9876543209876543,
1743
+ "grad_norm": 0.40206247568130493,
1744
+ "learning_rate": 8.966934456872602e-05,
1745
+ "loss": 0.0303,
1746
+ "step": 2480
1747
+ },
1748
+ {
1749
+ "epoch": 0.991636798088411,
1750
+ "grad_norm": 0.5897948741912842,
1751
+ "learning_rate": 8.95684782086195e-05,
1752
+ "loss": 0.0293,
1753
+ "step": 2490
1754
+ },
1755
+ {
1756
+ "epoch": 0.9956192751891677,
1757
+ "grad_norm": 0.3984488546848297,
1758
+ "learning_rate": 8.946717913397476e-05,
1759
+ "loss": 0.0374,
1760
+ "step": 2500
1761
+ },
1762
+ {
1763
+ "epoch": 0.9996017522899243,
1764
+ "grad_norm": 0.3679232597351074,
1765
+ "learning_rate": 8.93654484525822e-05,
1766
+ "loss": 0.0311,
1767
+ "step": 2510
1768
+ },
1769
+ {
1770
+ "epoch": 1.003584229390681,
1771
+ "grad_norm": 0.3514699637889862,
1772
+ "learning_rate": 8.926328727695226e-05,
1773
+ "loss": 0.0282,
1774
+ "step": 2520
1775
+ },
1776
+ {
1777
+ "epoch": 1.0075667064914378,
1778
+ "grad_norm": 0.34111514687538147,
1779
+ "learning_rate": 8.916069672430319e-05,
1780
+ "loss": 0.0272,
1781
+ "step": 2530
1782
+ },
1783
+ {
1784
+ "epoch": 1.0115491835921944,
1785
+ "grad_norm": 0.3726666569709778,
1786
+ "learning_rate": 8.905767791654884e-05,
1787
+ "loss": 0.0315,
1788
+ "step": 2540
1789
+ },
1790
+ {
1791
+ "epoch": 1.015531660692951,
1792
+ "grad_norm": 0.5761587619781494,
1793
+ "learning_rate": 8.895423198028638e-05,
1794
+ "loss": 0.0363,
1795
+ "step": 2550
1796
+ },
1797
+ {
1798
+ "epoch": 1.0195141377937076,
1799
+ "grad_norm": 0.5593805909156799,
1800
+ "learning_rate": 8.885036004678402e-05,
1801
+ "loss": 0.0311,
1802
+ "step": 2560
1803
+ },
1804
+ {
1805
+ "epoch": 1.0234966148944644,
1806
+ "grad_norm": 0.3937326669692993,
1807
+ "learning_rate": 8.874606325196857e-05,
1808
+ "loss": 0.0339,
1809
+ "step": 2570
1810
+ },
1811
+ {
1812
+ "epoch": 1.027479091995221,
1813
+ "grad_norm": 0.5169225931167603,
1814
+ "learning_rate": 8.864134273641304e-05,
1815
+ "loss": 0.03,
1816
+ "step": 2580
1817
+ },
1818
+ {
1819
+ "epoch": 1.0314615690959776,
1820
+ "grad_norm": 0.5744480490684509,
1821
+ "learning_rate": 8.853619964532427e-05,
1822
+ "loss": 0.0324,
1823
+ "step": 2590
1824
+ },
1825
+ {
1826
+ "epoch": 1.0354440461967345,
1827
+ "grad_norm": 0.555366039276123,
1828
+ "learning_rate": 8.843063512853019e-05,
1829
+ "loss": 0.0289,
1830
+ "step": 2600
1831
+ },
1832
+ {
1833
+ "epoch": 1.039426523297491,
1834
+ "grad_norm": 0.4377165734767914,
1835
+ "learning_rate": 8.832465034046749e-05,
1836
+ "loss": 0.0296,
1837
+ "step": 2610
1838
+ },
1839
+ {
1840
+ "epoch": 1.0434090003982477,
1841
+ "grad_norm": 0.5874027013778687,
1842
+ "learning_rate": 8.821824644016882e-05,
1843
+ "loss": 0.0317,
1844
+ "step": 2620
1845
+ },
1846
+ {
1847
+ "epoch": 1.0473914774990043,
1848
+ "grad_norm": 0.7405006289482117,
1849
+ "learning_rate": 8.811142459125019e-05,
1850
+ "loss": 0.0308,
1851
+ "step": 2630
1852
+ },
1853
+ {
1854
+ "epoch": 1.0513739545997611,
1855
+ "grad_norm": 0.4106619954109192,
1856
+ "learning_rate": 8.800418596189822e-05,
1857
+ "loss": 0.0252,
1858
+ "step": 2640
1859
+ },
1860
+ {
1861
+ "epoch": 1.0553564317005177,
1862
+ "grad_norm": 0.510388195514679,
1863
+ "learning_rate": 8.789653172485737e-05,
1864
+ "loss": 0.0297,
1865
+ "step": 2650
1866
+ },
1867
+ {
1868
+ "epoch": 1.0593389088012743,
1869
+ "grad_norm": 0.5874786972999573,
1870
+ "learning_rate": 8.778846305741715e-05,
1871
+ "loss": 0.0359,
1872
+ "step": 2660
1873
+ },
1874
+ {
1875
+ "epoch": 1.063321385902031,
1876
+ "grad_norm": 0.5348223447799683,
1877
+ "learning_rate": 8.767998114139918e-05,
1878
+ "loss": 0.0313,
1879
+ "step": 2670
1880
+ },
1881
+ {
1882
+ "epoch": 1.0673038630027878,
1883
+ "grad_norm": 0.3929119408130646,
1884
+ "learning_rate": 8.757108716314429e-05,
1885
+ "loss": 0.0329,
1886
+ "step": 2680
1887
+ },
1888
+ {
1889
+ "epoch": 1.0712863401035444,
1890
+ "grad_norm": 0.5612719058990479,
1891
+ "learning_rate": 8.746178231349962e-05,
1892
+ "loss": 0.0312,
1893
+ "step": 2690
1894
+ },
1895
+ {
1896
+ "epoch": 1.075268817204301,
1897
+ "grad_norm": 0.37800925970077515,
1898
+ "learning_rate": 8.735206778780549e-05,
1899
+ "loss": 0.0349,
1900
+ "step": 2700
1901
+ },
1902
+ {
1903
+ "epoch": 1.0792512943050578,
1904
+ "grad_norm": 0.43637755513191223,
1905
+ "learning_rate": 8.724194478588234e-05,
1906
+ "loss": 0.0307,
1907
+ "step": 2710
1908
+ },
1909
+ {
1910
+ "epoch": 1.0832337714058144,
1911
+ "grad_norm": 0.3608951270580292,
1912
+ "learning_rate": 8.713141451201772e-05,
1913
+ "loss": 0.0277,
1914
+ "step": 2720
1915
+ },
1916
+ {
1917
+ "epoch": 1.087216248506571,
1918
+ "grad_norm": 0.5240746140480042,
1919
+ "learning_rate": 8.702047817495295e-05,
1920
+ "loss": 0.0302,
1921
+ "step": 2730
1922
+ },
1923
+ {
1924
+ "epoch": 1.0911987256073277,
1925
+ "grad_norm": 0.3599265515804291,
1926
+ "learning_rate": 8.69091369878701e-05,
1927
+ "loss": 0.0313,
1928
+ "step": 2740
1929
+ },
1930
+ {
1931
+ "epoch": 1.0951812027080845,
1932
+ "grad_norm": 0.6075289845466614,
1933
+ "learning_rate": 8.679739216837849e-05,
1934
+ "loss": 0.0296,
1935
+ "step": 2750
1936
+ },
1937
+ {
1938
+ "epoch": 1.099163679808841,
1939
+ "grad_norm": 0.4289771318435669,
1940
+ "learning_rate": 8.66852449385016e-05,
1941
+ "loss": 0.0276,
1942
+ "step": 2760
1943
+ },
1944
+ {
1945
+ "epoch": 1.1031461569095977,
1946
+ "grad_norm": 0.3706158697605133,
1947
+ "learning_rate": 8.657269652466356e-05,
1948
+ "loss": 0.0278,
1949
+ "step": 2770
1950
+ },
1951
+ {
1952
+ "epoch": 1.1071286340103543,
1953
+ "grad_norm": 0.6201577186584473,
1954
+ "learning_rate": 8.645974815767577e-05,
1955
+ "loss": 0.0315,
1956
+ "step": 2780
1957
+ },
1958
+ {
1959
+ "epoch": 1.1111111111111112,
1960
+ "grad_norm": 0.5606251358985901,
1961
+ "learning_rate": 8.634640107272351e-05,
1962
+ "loss": 0.0274,
1963
+ "step": 2790
1964
+ },
1965
+ {
1966
+ "epoch": 1.1150935882118678,
1967
+ "grad_norm": 0.4653207063674927,
1968
+ "learning_rate": 8.623265650935234e-05,
1969
+ "loss": 0.0293,
1970
+ "step": 2800
1971
+ },
1972
+ {
1973
+ "epoch": 1.1190760653126244,
1974
+ "grad_norm": 0.4102509915828705,
1975
+ "learning_rate": 8.611851571145456e-05,
1976
+ "loss": 0.0311,
1977
+ "step": 2810
1978
+ },
1979
+ {
1980
+ "epoch": 1.1230585424133812,
1981
+ "grad_norm": 0.5217996835708618,
1982
+ "learning_rate": 8.600397992725566e-05,
1983
+ "loss": 0.0248,
1984
+ "step": 2820
1985
+ },
1986
+ {
1987
+ "epoch": 1.1270410195141378,
1988
+ "grad_norm": 0.5519301295280457,
1989
+ "learning_rate": 8.588905040930061e-05,
1990
+ "loss": 0.0314,
1991
+ "step": 2830
1992
+ },
1993
+ {
1994
+ "epoch": 1.1310234966148944,
1995
+ "grad_norm": 0.43869367241859436,
1996
+ "learning_rate": 8.577372841444022e-05,
1997
+ "loss": 0.0289,
1998
+ "step": 2840
1999
+ },
2000
+ {
2001
+ "epoch": 1.135005973715651,
2002
+ "grad_norm": 0.4673241078853607,
2003
+ "learning_rate": 8.565801520381736e-05,
2004
+ "loss": 0.026,
2005
+ "step": 2850
2006
+ },
2007
+ {
2008
+ "epoch": 1.1389884508164079,
2009
+ "grad_norm": 0.4510667622089386,
2010
+ "learning_rate": 8.554191204285313e-05,
2011
+ "loss": 0.0319,
2012
+ "step": 2860
2013
+ },
2014
+ {
2015
+ "epoch": 1.1429709279171645,
2016
+ "grad_norm": 0.473105788230896,
2017
+ "learning_rate": 8.542542020123315e-05,
2018
+ "loss": 0.032,
2019
+ "step": 2870
2020
+ },
2021
+ {
2022
+ "epoch": 1.146953405017921,
2023
+ "grad_norm": 0.4328536093235016,
2024
+ "learning_rate": 8.530854095289347e-05,
2025
+ "loss": 0.029,
2026
+ "step": 2880
2027
+ },
2028
+ {
2029
+ "epoch": 1.1509358821186777,
2030
+ "grad_norm": 0.45955684781074524,
2031
+ "learning_rate": 8.519127557600688e-05,
2032
+ "loss": 0.0274,
2033
+ "step": 2890
2034
+ },
2035
+ {
2036
+ "epoch": 1.1549183592194345,
2037
+ "grad_norm": 0.3034651577472687,
2038
+ "learning_rate": 8.507362535296871e-05,
2039
+ "loss": 0.0286,
2040
+ "step": 2900
2041
+ },
2042
+ {
2043
+ "epoch": 1.1589008363201911,
2044
+ "grad_norm": 0.6441288590431213,
2045
+ "learning_rate": 8.495559157038299e-05,
2046
+ "loss": 0.0377,
2047
+ "step": 2910
2048
+ },
2049
+ {
2050
+ "epoch": 1.1628833134209478,
2051
+ "grad_norm": 0.43017473816871643,
2052
+ "learning_rate": 8.483717551904823e-05,
2053
+ "loss": 0.0297,
2054
+ "step": 2920
2055
+ },
2056
+ {
2057
+ "epoch": 1.1668657905217046,
2058
+ "grad_norm": 0.5794181227684021,
2059
+ "learning_rate": 8.47183784939434e-05,
2060
+ "loss": 0.0281,
2061
+ "step": 2930
2062
+ },
2063
+ {
2064
+ "epoch": 1.1708482676224612,
2065
+ "grad_norm": 0.32613155245780945,
2066
+ "learning_rate": 8.459920179421374e-05,
2067
+ "loss": 0.0324,
2068
+ "step": 2940
2069
+ },
2070
+ {
2071
+ "epoch": 1.1748307447232178,
2072
+ "grad_norm": 0.514627993106842,
2073
+ "learning_rate": 8.447964672315656e-05,
2074
+ "loss": 0.0275,
2075
+ "step": 2950
2076
+ },
2077
+ {
2078
+ "epoch": 1.1788132218239746,
2079
+ "grad_norm": 0.30379730463027954,
2080
+ "learning_rate": 8.435971458820692e-05,
2081
+ "loss": 0.0322,
2082
+ "step": 2960
2083
+ },
2084
+ {
2085
+ "epoch": 1.1827956989247312,
2086
+ "grad_norm": 0.5878794193267822,
2087
+ "learning_rate": 8.423940670092345e-05,
2088
+ "loss": 0.0318,
2089
+ "step": 2970
2090
+ },
2091
+ {
2092
+ "epoch": 1.1867781760254879,
2093
+ "grad_norm": 0.3879925012588501,
2094
+ "learning_rate": 8.411872437697394e-05,
2095
+ "loss": 0.0323,
2096
+ "step": 2980
2097
+ },
2098
+ {
2099
+ "epoch": 1.1907606531262445,
2100
+ "grad_norm": 0.2640171945095062,
2101
+ "learning_rate": 8.399766893612096e-05,
2102
+ "loss": 0.0265,
2103
+ "step": 2990
2104
+ },
2105
+ {
2106
+ "epoch": 1.194743130227001,
2107
+ "grad_norm": 0.4299219846725464,
2108
+ "learning_rate": 8.38762417022074e-05,
2109
+ "loss": 0.0312,
2110
+ "step": 3000
2111
+ },
2112
+ {
2113
+ "epoch": 1.198725607327758,
2114
+ "grad_norm": 0.3073621690273285,
2115
+ "learning_rate": 8.375444400314204e-05,
2116
+ "loss": 0.0253,
2117
+ "step": 3010
2118
+ },
2119
+ {
2120
+ "epoch": 1.2027080844285145,
2121
+ "grad_norm": 0.26146814227104187,
2122
+ "learning_rate": 8.3632277170885e-05,
2123
+ "loss": 0.0312,
2124
+ "step": 3020
2125
+ },
2126
+ {
2127
+ "epoch": 1.2066905615292711,
2128
+ "grad_norm": 0.2975267767906189,
2129
+ "learning_rate": 8.350974254143318e-05,
2130
+ "loss": 0.0271,
2131
+ "step": 3030
2132
+ },
2133
+ {
2134
+ "epoch": 1.210673038630028,
2135
+ "grad_norm": 0.20422956347465515,
2136
+ "learning_rate": 8.338684145480566e-05,
2137
+ "loss": 0.0259,
2138
+ "step": 3040
2139
+ },
2140
+ {
2141
+ "epoch": 1.2146555157307846,
2142
+ "grad_norm": 0.31191202998161316,
2143
+ "learning_rate": 8.326357525502904e-05,
2144
+ "loss": 0.0304,
2145
+ "step": 3050
2146
+ },
2147
+ {
2148
+ "epoch": 1.2186379928315412,
2149
+ "grad_norm": 0.5126979947090149,
2150
+ "learning_rate": 8.313994529012273e-05,
2151
+ "loss": 0.0294,
2152
+ "step": 3060
2153
+ },
2154
+ {
2155
+ "epoch": 1.222620469932298,
2156
+ "grad_norm": 0.3219705820083618,
2157
+ "learning_rate": 8.301595291208422e-05,
2158
+ "loss": 0.0305,
2159
+ "step": 3070
2160
+ },
2161
+ {
2162
+ "epoch": 1.2266029470330546,
2163
+ "grad_norm": 0.4278619587421417,
2164
+ "learning_rate": 8.289159947687427e-05,
2165
+ "loss": 0.0297,
2166
+ "step": 3080
2167
+ },
2168
+ {
2169
+ "epoch": 1.2305854241338112,
2170
+ "grad_norm": 0.5774158239364624,
2171
+ "learning_rate": 8.276688634440216e-05,
2172
+ "loss": 0.0296,
2173
+ "step": 3090
2174
+ },
2175
+ {
2176
+ "epoch": 1.2345679012345678,
2177
+ "grad_norm": 0.5146026015281677,
2178
+ "learning_rate": 8.26418148785107e-05,
2179
+ "loss": 0.0311,
2180
+ "step": 3100
2181
+ },
2182
+ {
2183
+ "epoch": 1.2385503783353247,
2184
+ "grad_norm": 0.347504198551178,
2185
+ "learning_rate": 8.251638644696141e-05,
2186
+ "loss": 0.0345,
2187
+ "step": 3110
2188
+ },
2189
+ {
2190
+ "epoch": 1.2425328554360813,
2191
+ "grad_norm": 0.5153434872627258,
2192
+ "learning_rate": 8.23906024214195e-05,
2193
+ "loss": 0.0276,
2194
+ "step": 3120
2195
+ },
2196
+ {
2197
+ "epoch": 1.2465153325368379,
2198
+ "grad_norm": 0.31597134470939636,
2199
+ "learning_rate": 8.226446417743897e-05,
2200
+ "loss": 0.0343,
2201
+ "step": 3130
2202
+ },
2203
+ {
2204
+ "epoch": 1.2504978096375945,
2205
+ "grad_norm": 0.4005573093891144,
2206
+ "learning_rate": 8.213797309444742e-05,
2207
+ "loss": 0.0294,
2208
+ "step": 3140
2209
+ },
2210
+ {
2211
+ "epoch": 1.2544802867383513,
2212
+ "grad_norm": 0.3009648025035858,
2213
+ "learning_rate": 8.201113055573105e-05,
2214
+ "loss": 0.0287,
2215
+ "step": 3150
2216
+ },
2217
+ {
2218
+ "epoch": 1.258462763839108,
2219
+ "grad_norm": 0.6014096140861511,
2220
+ "learning_rate": 8.188393794841958e-05,
2221
+ "loss": 0.0261,
2222
+ "step": 3160
2223
+ },
2224
+ {
2225
+ "epoch": 1.2624452409398645,
2226
+ "grad_norm": 0.38871604204177856,
2227
+ "learning_rate": 8.175639666347094e-05,
2228
+ "loss": 0.0316,
2229
+ "step": 3170
2230
+ },
2231
+ {
2232
+ "epoch": 1.2664277180406214,
2233
+ "grad_norm": 0.4298296570777893,
2234
+ "learning_rate": 8.162850809565623e-05,
2235
+ "loss": 0.0248,
2236
+ "step": 3180
2237
+ },
2238
+ {
2239
+ "epoch": 1.270410195141378,
2240
+ "grad_norm": 0.3726460039615631,
2241
+ "learning_rate": 8.150027364354431e-05,
2242
+ "loss": 0.0312,
2243
+ "step": 3190
2244
+ },
2245
+ {
2246
+ "epoch": 1.2743926722421346,
2247
+ "grad_norm": 0.5679003000259399,
2248
+ "learning_rate": 8.137169470948662e-05,
2249
+ "loss": 0.0346,
2250
+ "step": 3200
2251
+ },
2252
+ {
2253
+ "epoch": 1.2783751493428912,
2254
+ "grad_norm": 0.3676661252975464,
2255
+ "learning_rate": 8.124277269960179e-05,
2256
+ "loss": 0.0298,
2257
+ "step": 3210
2258
+ },
2259
+ {
2260
+ "epoch": 1.2823576264436478,
2261
+ "grad_norm": 0.34625566005706787,
2262
+ "learning_rate": 8.111350902376023e-05,
2263
+ "loss": 0.0276,
2264
+ "step": 3220
2265
+ },
2266
+ {
2267
+ "epoch": 1.2863401035444046,
2268
+ "grad_norm": 0.37914222478866577,
2269
+ "learning_rate": 8.098390509556883e-05,
2270
+ "loss": 0.028,
2271
+ "step": 3230
2272
+ },
2273
+ {
2274
+ "epoch": 1.2903225806451613,
2275
+ "grad_norm": 0.6115908622741699,
2276
+ "learning_rate": 8.085396233235536e-05,
2277
+ "loss": 0.027,
2278
+ "step": 3240
2279
+ },
2280
+ {
2281
+ "epoch": 1.2943050577459179,
2282
+ "grad_norm": 0.3943430781364441,
2283
+ "learning_rate": 8.072368215515306e-05,
2284
+ "loss": 0.0264,
2285
+ "step": 3250
2286
+ },
2287
+ {
2288
+ "epoch": 1.2982875348466747,
2289
+ "grad_norm": 0.4377286434173584,
2290
+ "learning_rate": 8.059306598868506e-05,
2291
+ "loss": 0.0334,
2292
+ "step": 3260
2293
+ },
2294
+ {
2295
+ "epoch": 1.3022700119474313,
2296
+ "grad_norm": 0.28060171008110046,
2297
+ "learning_rate": 8.046211526134888e-05,
2298
+ "loss": 0.0284,
2299
+ "step": 3270
2300
+ },
2301
+ {
2302
+ "epoch": 1.306252489048188,
2303
+ "grad_norm": 0.4140307605266571,
2304
+ "learning_rate": 8.033083140520065e-05,
2305
+ "loss": 0.027,
2306
+ "step": 3280
2307
+ },
2308
+ {
2309
+ "epoch": 1.3102349661489447,
2310
+ "grad_norm": 0.334888756275177,
2311
+ "learning_rate": 8.019921585593962e-05,
2312
+ "loss": 0.0295,
2313
+ "step": 3290
2314
+ },
2315
+ {
2316
+ "epoch": 1.3142174432497014,
2317
+ "grad_norm": 0.5044397115707397,
2318
+ "learning_rate": 8.006727005289232e-05,
2319
+ "loss": 0.0239,
2320
+ "step": 3300
2321
+ },
2322
+ {
2323
+ "epoch": 1.318199920350458,
2324
+ "grad_norm": 0.39227598905563354,
2325
+ "learning_rate": 7.993499543899692e-05,
2326
+ "loss": 0.029,
2327
+ "step": 3310
2328
+ },
2329
+ {
2330
+ "epoch": 1.3221823974512146,
2331
+ "grad_norm": 0.5009192824363708,
2332
+ "learning_rate": 7.980239346078742e-05,
2333
+ "loss": 0.0276,
2334
+ "step": 3320
2335
+ },
2336
+ {
2337
+ "epoch": 1.3261648745519714,
2338
+ "grad_norm": 0.33442339301109314,
2339
+ "learning_rate": 7.966946556837778e-05,
2340
+ "loss": 0.025,
2341
+ "step": 3330
2342
+ },
2343
+ {
2344
+ "epoch": 1.330147351652728,
2345
+ "grad_norm": 0.44179999828338623,
2346
+ "learning_rate": 7.953621321544616e-05,
2347
+ "loss": 0.0292,
2348
+ "step": 3340
2349
+ },
2350
+ {
2351
+ "epoch": 1.3341298287534846,
2352
+ "grad_norm": 0.48963797092437744,
2353
+ "learning_rate": 7.940263785921896e-05,
2354
+ "loss": 0.0287,
2355
+ "step": 3350
2356
+ },
2357
+ {
2358
+ "epoch": 1.3381123058542412,
2359
+ "grad_norm": 0.35747164487838745,
2360
+ "learning_rate": 7.926874096045482e-05,
2361
+ "loss": 0.0257,
2362
+ "step": 3360
2363
+ },
2364
+ {
2365
+ "epoch": 1.342094782954998,
2366
+ "grad_norm": 0.6534529328346252,
2367
+ "learning_rate": 7.913452398342881e-05,
2368
+ "loss": 0.036,
2369
+ "step": 3370
2370
+ },
2371
+ {
2372
+ "epoch": 1.3460772600557547,
2373
+ "grad_norm": 0.4266543686389923,
2374
+ "learning_rate": 7.89999883959163e-05,
2375
+ "loss": 0.0307,
2376
+ "step": 3380
2377
+ },
2378
+ {
2379
+ "epoch": 1.3500597371565113,
2380
+ "grad_norm": 0.5537866950035095,
2381
+ "learning_rate": 7.886513566917687e-05,
2382
+ "loss": 0.0307,
2383
+ "step": 3390
2384
+ },
2385
+ {
2386
+ "epoch": 1.3540422142572681,
2387
+ "grad_norm": 0.35446301102638245,
2388
+ "learning_rate": 7.872996727793838e-05,
2389
+ "loss": 0.0323,
2390
+ "step": 3400
2391
+ },
2392
+ {
2393
+ "epoch": 1.3580246913580247,
2394
+ "grad_norm": 0.4526813328266144,
2395
+ "learning_rate": 7.859448470038069e-05,
2396
+ "loss": 0.0275,
2397
+ "step": 3410
2398
+ },
2399
+ {
2400
+ "epoch": 1.3620071684587813,
2401
+ "grad_norm": 0.49604639410972595,
2402
+ "learning_rate": 7.845868941811956e-05,
2403
+ "loss": 0.0317,
2404
+ "step": 3420
2405
+ },
2406
+ {
2407
+ "epoch": 1.3659896455595382,
2408
+ "grad_norm": 0.3144626319408417,
2409
+ "learning_rate": 7.832258291619043e-05,
2410
+ "loss": 0.0276,
2411
+ "step": 3430
2412
+ },
2413
+ {
2414
+ "epoch": 1.3699721226602948,
2415
+ "grad_norm": 0.5718449354171753,
2416
+ "learning_rate": 7.81861666830322e-05,
2417
+ "loss": 0.0261,
2418
+ "step": 3440
2419
+ },
2420
+ {
2421
+ "epoch": 1.3739545997610514,
2422
+ "grad_norm": 0.3567497134208679,
2423
+ "learning_rate": 7.804944221047097e-05,
2424
+ "loss": 0.027,
2425
+ "step": 3450
2426
+ },
2427
+ {
2428
+ "epoch": 1.377937076861808,
2429
+ "grad_norm": 0.5491267442703247,
2430
+ "learning_rate": 7.791241099370364e-05,
2431
+ "loss": 0.033,
2432
+ "step": 3460
2433
+ },
2434
+ {
2435
+ "epoch": 1.3819195539625646,
2436
+ "grad_norm": 0.46909114718437195,
2437
+ "learning_rate": 7.777507453128163e-05,
2438
+ "loss": 0.0298,
2439
+ "step": 3470
2440
+ },
2441
+ {
2442
+ "epoch": 1.3859020310633214,
2443
+ "grad_norm": 0.4308125376701355,
2444
+ "learning_rate": 7.763743432509451e-05,
2445
+ "loss": 0.0278,
2446
+ "step": 3480
2447
+ },
2448
+ {
2449
+ "epoch": 1.389884508164078,
2450
+ "grad_norm": 0.4042797088623047,
2451
+ "learning_rate": 7.749949188035353e-05,
2452
+ "loss": 0.0312,
2453
+ "step": 3490
2454
+ },
2455
+ {
2456
+ "epoch": 1.3938669852648347,
2457
+ "grad_norm": 0.4012243151664734,
2458
+ "learning_rate": 7.736124870557516e-05,
2459
+ "loss": 0.0296,
2460
+ "step": 3500
2461
+ },
2462
+ {
2463
+ "epoch": 1.3978494623655915,
2464
+ "grad_norm": 0.31076422333717346,
2465
+ "learning_rate": 7.722270631256459e-05,
2466
+ "loss": 0.0306,
2467
+ "step": 3510
2468
+ },
2469
+ {
2470
+ "epoch": 1.401831939466348,
2471
+ "grad_norm": 0.4105532467365265,
2472
+ "learning_rate": 7.708386621639925e-05,
2473
+ "loss": 0.0262,
2474
+ "step": 3520
2475
+ },
2476
+ {
2477
+ "epoch": 1.4058144165671047,
2478
+ "grad_norm": 0.4460810124874115,
2479
+ "learning_rate": 7.694472993541219e-05,
2480
+ "loss": 0.0269,
2481
+ "step": 3530
2482
+ },
2483
+ {
2484
+ "epoch": 1.4097968936678615,
2485
+ "grad_norm": 0.4327501952648163,
2486
+ "learning_rate": 7.680529899117547e-05,
2487
+ "loss": 0.0313,
2488
+ "step": 3540
2489
+ },
2490
+ {
2491
+ "epoch": 1.4137793707686181,
2492
+ "grad_norm": 0.4426930248737335,
2493
+ "learning_rate": 7.666557490848358e-05,
2494
+ "loss": 0.0289,
2495
+ "step": 3550
2496
+ },
2497
+ {
2498
+ "epoch": 1.4177618478693748,
2499
+ "grad_norm": 0.4433678388595581,
2500
+ "learning_rate": 7.65255592153367e-05,
2501
+ "loss": 0.0307,
2502
+ "step": 3560
2503
+ },
2504
+ {
2505
+ "epoch": 1.4217443249701314,
2506
+ "grad_norm": 0.39277946949005127,
2507
+ "learning_rate": 7.638525344292402e-05,
2508
+ "loss": 0.0321,
2509
+ "step": 3570
2510
+ },
2511
+ {
2512
+ "epoch": 1.425726802070888,
2513
+ "grad_norm": 0.4349759817123413,
2514
+ "learning_rate": 7.624465912560697e-05,
2515
+ "loss": 0.0319,
2516
+ "step": 3580
2517
+ },
2518
+ {
2519
+ "epoch": 1.4297092791716448,
2520
+ "grad_norm": 0.36243054270744324,
2521
+ "learning_rate": 7.610377780090249e-05,
2522
+ "loss": 0.025,
2523
+ "step": 3590
2524
+ },
2525
+ {
2526
+ "epoch": 1.4336917562724014,
2527
+ "grad_norm": 0.23596417903900146,
2528
+ "learning_rate": 7.596261100946618e-05,
2529
+ "loss": 0.0274,
2530
+ "step": 3600
2531
+ },
2532
+ {
2533
+ "epoch": 1.437674233373158,
2534
+ "grad_norm": 0.3828240633010864,
2535
+ "learning_rate": 7.582116029507542e-05,
2536
+ "loss": 0.0259,
2537
+ "step": 3610
2538
+ },
2539
+ {
2540
+ "epoch": 1.4416567104739149,
2541
+ "grad_norm": 0.4790317714214325,
2542
+ "learning_rate": 7.56794272046126e-05,
2543
+ "loss": 0.0278,
2544
+ "step": 3620
2545
+ },
2546
+ {
2547
+ "epoch": 1.4456391875746715,
2548
+ "grad_norm": 0.46134695410728455,
2549
+ "learning_rate": 7.55374132880481e-05,
2550
+ "loss": 0.0298,
2551
+ "step": 3630
2552
+ },
2553
+ {
2554
+ "epoch": 1.449621664675428,
2555
+ "grad_norm": 0.4513108730316162,
2556
+ "learning_rate": 7.539512009842333e-05,
2557
+ "loss": 0.0313,
2558
+ "step": 3640
2559
+ },
2560
+ {
2561
+ "epoch": 1.453604141776185,
2562
+ "grad_norm": 0.5272055268287659,
2563
+ "learning_rate": 7.525254919183382e-05,
2564
+ "loss": 0.0283,
2565
+ "step": 3650
2566
+ },
2567
+ {
2568
+ "epoch": 1.4575866188769415,
2569
+ "grad_norm": 0.43485143780708313,
2570
+ "learning_rate": 7.510970212741215e-05,
2571
+ "loss": 0.0272,
2572
+ "step": 3660
2573
+ },
2574
+ {
2575
+ "epoch": 1.4615690959776981,
2576
+ "grad_norm": 0.44522616267204285,
2577
+ "learning_rate": 7.496658046731096e-05,
2578
+ "loss": 0.0281,
2579
+ "step": 3670
2580
+ },
2581
+ {
2582
+ "epoch": 1.4655515730784547,
2583
+ "grad_norm": 0.7030194997787476,
2584
+ "learning_rate": 7.482318577668578e-05,
2585
+ "loss": 0.0336,
2586
+ "step": 3680
2587
+ },
2588
+ {
2589
+ "epoch": 1.4695340501792113,
2590
+ "grad_norm": 0.47834673523902893,
2591
+ "learning_rate": 7.467951962367796e-05,
2592
+ "loss": 0.0276,
2593
+ "step": 3690
2594
+ },
2595
+ {
2596
+ "epoch": 1.4735165272799682,
2597
+ "grad_norm": 0.4326513409614563,
2598
+ "learning_rate": 7.453558357939755e-05,
2599
+ "loss": 0.0268,
2600
+ "step": 3700
2601
+ },
2602
+ {
2603
+ "epoch": 1.4774990043807248,
2604
+ "grad_norm": 0.4779656231403351,
2605
+ "learning_rate": 7.439137921790606e-05,
2606
+ "loss": 0.0322,
2607
+ "step": 3710
2608
+ },
2609
+ {
2610
+ "epoch": 1.4814814814814814,
2611
+ "grad_norm": 0.5998985767364502,
2612
+ "learning_rate": 7.42469081161993e-05,
2613
+ "loss": 0.0257,
2614
+ "step": 3720
2615
+ },
2616
+ {
2617
+ "epoch": 1.4854639585822382,
2618
+ "grad_norm": 0.44799843430519104,
2619
+ "learning_rate": 7.410217185419006e-05,
2620
+ "loss": 0.0235,
2621
+ "step": 3730
2622
+ },
2623
+ {
2624
+ "epoch": 1.4894464356829948,
2625
+ "grad_norm": 0.335964560508728,
2626
+ "learning_rate": 7.395717201469095e-05,
2627
+ "loss": 0.0286,
2628
+ "step": 3740
2629
+ },
2630
+ {
2631
+ "epoch": 1.4934289127837514,
2632
+ "grad_norm": 0.5583806037902832,
2633
+ "learning_rate": 7.381191018339696e-05,
2634
+ "loss": 0.0292,
2635
+ "step": 3750
2636
+ },
2637
+ {
2638
+ "epoch": 1.4974113898845083,
2639
+ "grad_norm": 0.40853527188301086,
2640
+ "learning_rate": 7.36663879488682e-05,
2641
+ "loss": 0.024,
2642
+ "step": 3760
2643
+ },
2644
+ {
2645
+ "epoch": 1.5013938669852647,
2646
+ "grad_norm": 0.5366830825805664,
2647
+ "learning_rate": 7.352060690251254e-05,
2648
+ "loss": 0.0278,
2649
+ "step": 3770
2650
+ },
2651
+ {
2652
+ "epoch": 1.5053763440860215,
2653
+ "grad_norm": 0.5075726509094238,
2654
+ "learning_rate": 7.337456863856811e-05,
2655
+ "loss": 0.031,
2656
+ "step": 3780
2657
+ },
2658
+ {
2659
+ "epoch": 1.5093588211867783,
2660
+ "grad_norm": 0.4533558487892151,
2661
+ "learning_rate": 7.3228274754086e-05,
2662
+ "loss": 0.0232,
2663
+ "step": 3790
2664
+ },
2665
+ {
2666
+ "epoch": 1.5133412982875347,
2667
+ "grad_norm": 0.4715867340564728,
2668
+ "learning_rate": 7.308172684891267e-05,
2669
+ "loss": 0.0319,
2670
+ "step": 3800
2671
+ },
2672
+ {
2673
+ "epoch": 1.5173237753882916,
2674
+ "grad_norm": 0.2563737630844116,
2675
+ "learning_rate": 7.293492652567255e-05,
2676
+ "loss": 0.0239,
2677
+ "step": 3810
2678
+ },
2679
+ {
2680
+ "epoch": 1.5213062524890482,
2681
+ "grad_norm": 0.3013695478439331,
2682
+ "learning_rate": 7.278787538975043e-05,
2683
+ "loss": 0.0302,
2684
+ "step": 3820
2685
+ },
2686
+ {
2687
+ "epoch": 1.5252887295898048,
2688
+ "grad_norm": 0.39960938692092896,
2689
+ "learning_rate": 7.2640575049274e-05,
2690
+ "loss": 0.0251,
2691
+ "step": 3830
2692
+ },
2693
+ {
2694
+ "epoch": 1.5292712066905616,
2695
+ "grad_norm": 0.380015105009079,
2696
+ "learning_rate": 7.249302711509616e-05,
2697
+ "loss": 0.0231,
2698
+ "step": 3840
2699
+ },
2700
+ {
2701
+ "epoch": 1.5332536837913182,
2702
+ "grad_norm": 0.2648374140262604,
2703
+ "learning_rate": 7.23452332007775e-05,
2704
+ "loss": 0.0225,
2705
+ "step": 3850
2706
+ },
2707
+ {
2708
+ "epoch": 1.5372361608920748,
2709
+ "grad_norm": 0.32741981744766235,
2710
+ "learning_rate": 7.219719492256858e-05,
2711
+ "loss": 0.0272,
2712
+ "step": 3860
2713
+ },
2714
+ {
2715
+ "epoch": 1.5412186379928317,
2716
+ "grad_norm": 0.5281572341918945,
2717
+ "learning_rate": 7.20489138993923e-05,
2718
+ "loss": 0.025,
2719
+ "step": 3870
2720
+ },
2721
+ {
2722
+ "epoch": 1.5452011150935883,
2723
+ "grad_norm": 0.3067561388015747,
2724
+ "learning_rate": 7.190039175282614e-05,
2725
+ "loss": 0.0261,
2726
+ "step": 3880
2727
+ },
2728
+ {
2729
+ "epoch": 1.5491835921943449,
2730
+ "grad_norm": 0.4421009421348572,
2731
+ "learning_rate": 7.175163010708455e-05,
2732
+ "loss": 0.0245,
2733
+ "step": 3890
2734
+ },
2735
+ {
2736
+ "epoch": 1.5531660692951017,
2737
+ "grad_norm": 0.3433239758014679,
2738
+ "learning_rate": 7.1602630589001e-05,
2739
+ "loss": 0.027,
2740
+ "step": 3900
2741
+ },
2742
+ {
2743
+ "epoch": 1.557148546395858,
2744
+ "grad_norm": 0.30050766468048096,
2745
+ "learning_rate": 7.14533948280104e-05,
2746
+ "loss": 0.0269,
2747
+ "step": 3910
2748
+ },
2749
+ {
2750
+ "epoch": 1.561131023496615,
2751
+ "grad_norm": 0.3128851056098938,
2752
+ "learning_rate": 7.130392445613109e-05,
2753
+ "loss": 0.03,
2754
+ "step": 3920
2755
+ },
2756
+ {
2757
+ "epoch": 1.5651135005973715,
2758
+ "grad_norm": 0.3643544912338257,
2759
+ "learning_rate": 7.115422110794711e-05,
2760
+ "loss": 0.0278,
2761
+ "step": 3930
2762
+ },
2763
+ {
2764
+ "epoch": 1.5690959776981281,
2765
+ "grad_norm": 0.4426794648170471,
2766
+ "learning_rate": 7.100428642059033e-05,
2767
+ "loss": 0.0303,
2768
+ "step": 3940
2769
+ },
2770
+ {
2771
+ "epoch": 1.573078454798885,
2772
+ "grad_norm": 0.3874496817588806,
2773
+ "learning_rate": 7.08541220337224e-05,
2774
+ "loss": 0.0277,
2775
+ "step": 3950
2776
+ },
2777
+ {
2778
+ "epoch": 1.5770609318996416,
2779
+ "grad_norm": 0.3397146761417389,
2780
+ "learning_rate": 7.070372958951706e-05,
2781
+ "loss": 0.0271,
2782
+ "step": 3960
2783
+ },
2784
+ {
2785
+ "epoch": 1.5810434090003982,
2786
+ "grad_norm": 0.6062551736831665,
2787
+ "learning_rate": 7.055311073264194e-05,
2788
+ "loss": 0.0334,
2789
+ "step": 3970
2790
+ },
2791
+ {
2792
+ "epoch": 1.585025886101155,
2793
+ "grad_norm": 0.2893720269203186,
2794
+ "learning_rate": 7.040226711024077e-05,
2795
+ "loss": 0.0276,
2796
+ "step": 3980
2797
+ },
2798
+ {
2799
+ "epoch": 1.5890083632019116,
2800
+ "grad_norm": 0.28936243057250977,
2801
+ "learning_rate": 7.02512003719152e-05,
2802
+ "loss": 0.0239,
2803
+ "step": 3990
2804
+ },
2805
+ {
2806
+ "epoch": 1.5929908403026682,
2807
+ "grad_norm": 0.4626624286174774,
2808
+ "learning_rate": 7.00999121697069e-05,
2809
+ "loss": 0.0304,
2810
+ "step": 4000
2811
+ }
2812
+ ],
2813
+ "logging_steps": 10,
2814
+ "max_steps": 10000,
2815
+ "num_input_tokens_seen": 0,
2816
+ "num_train_epochs": 4,
2817
+ "save_steps": 1000,
2818
+ "stateful_callbacks": {
2819
+ "TrainerControl": {
2820
+ "args": {
2821
+ "should_epoch_stop": false,
2822
+ "should_evaluate": false,
2823
+ "should_log": false,
2824
+ "should_save": true,
2825
+ "should_training_stop": false
2826
+ },
2827
+ "attributes": {}
2828
+ }
2829
+ },
2830
+ "total_flos": 0.0,
2831
+ "train_batch_size": 32,
2832
+ "trial_name": null,
2833
+ "trial_params": null
2834
+ }
checkpoint-5000/config.json ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "action_dim": 32,
3
+ "action_head_cfg": {
4
+ "action_dim": 32,
5
+ "action_horizon": 16,
6
+ "add_pos_embed": true,
7
+ "backbone_embedding_dim": 2048,
8
+ "diffusion_model_cfg": {
9
+ "attention_head_dim": 48,
10
+ "cross_attention_dim": 2048,
11
+ "dropout": 0.2,
12
+ "final_dropout": true,
13
+ "interleave_self_attention": true,
14
+ "norm_type": "ada_norm",
15
+ "num_attention_heads": 32,
16
+ "num_layers": 16,
17
+ "output_dim": 1024,
18
+ "positional_embeddings": null
19
+ },
20
+ "hidden_size": 1024,
21
+ "input_embedding_dim": 1536,
22
+ "max_action_dim": 32,
23
+ "max_state_dim": 64,
24
+ "model_dtype": "float32",
25
+ "noise_beta_alpha": 1.5,
26
+ "noise_beta_beta": 1.0,
27
+ "noise_s": 0.999,
28
+ "num_inference_timesteps": 4,
29
+ "num_target_vision_tokens": 32,
30
+ "num_timestep_buckets": 1000,
31
+ "tune_diffusion_model": true,
32
+ "tune_projector": true,
33
+ "use_vlln": true,
34
+ "vl_self_attention_cfg": {
35
+ "attention_head_dim": 64,
36
+ "dropout": 0.2,
37
+ "final_dropout": true,
38
+ "num_attention_heads": 32,
39
+ "num_layers": 4,
40
+ "positional_embeddings": null
41
+ }
42
+ },
43
+ "action_horizon": 16,
44
+ "architectures": [
45
+ "GR00T_N1_5"
46
+ ],
47
+ "attn_implementation": null,
48
+ "backbone_cfg": {
49
+ "eagle_path": "NVEagle/eagle_er-qwen3_1_7B-Siglip2_400M_stage1_5_128gpu_er_v7_1mlp_nops",
50
+ "load_bf16": false,
51
+ "project_to_dim": null,
52
+ "reproject_vision": false,
53
+ "select_layer": 12,
54
+ "tune_llm": false,
55
+ "tune_visual": true,
56
+ "use_flash_attention": true
57
+ },
58
+ "compute_dtype": "bfloat16",
59
+ "hidden_size": 2048,
60
+ "model_dtype": "float32",
61
+ "model_type": "gr00t_n1_5",
62
+ "torch_dtype": "bfloat16",
63
+ "transformers_version": "4.51.3"
64
+ }
checkpoint-5000/experiment_cfg/metadata.json ADDED
@@ -0,0 +1,195 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "new_embodiment": {
3
+ "statistics": {
4
+ "state": {
5
+ "single_arm": {
6
+ "max": [
7
+ 33.22265625,
8
+ 86.1328125,
9
+ 84.7265625,
10
+ 101.337890625,
11
+ 32.783203125
12
+ ],
13
+ "min": [
14
+ -87.275390625,
15
+ -103.974609375,
16
+ -106.435546875,
17
+ -16.34765625,
18
+ -170.068359375
19
+ ],
20
+ "mean": [
21
+ -26.84341812133789,
22
+ 6.497307777404785,
23
+ -22.14896583557129,
24
+ 74.85757446289062,
25
+ -111.02351379394531
26
+ ],
27
+ "std": [
28
+ 27.50257110595703,
29
+ 41.151065826416016,
30
+ 48.41417694091797,
31
+ 20.208982467651367,
32
+ 46.13625717163086
33
+ ],
34
+ "q01": [
35
+ -79.013671875,
36
+ -103.271484375,
37
+ -104.0625,
38
+ 19.187402343750005,
39
+ -169.892578125
40
+ ],
41
+ "q99": [
42
+ 24.169921875,
43
+ 73.30078125,
44
+ 84.375,
45
+ 101.07421875,
46
+ 11.162109375
47
+ ]
48
+ },
49
+ "gripper": {
50
+ "max": [
51
+ 92.373046875
52
+ ],
53
+ "min": [
54
+ -0.615234375
55
+ ],
56
+ "mean": [
57
+ 14.521344184875488
58
+ ],
59
+ "std": [
60
+ 13.066176414489746
61
+ ],
62
+ "q01": [
63
+ 0.0
64
+ ],
65
+ "q99": [
66
+ 52.470703125
67
+ ]
68
+ }
69
+ },
70
+ "action": {
71
+ "single_arm": {
72
+ "max": [
73
+ 33.662109375,
74
+ 86.484375,
75
+ 85.078125,
76
+ 102.65625,
77
+ 33.3984375
78
+ ],
79
+ "min": [
80
+ -87.890625,
81
+ -104.4140625,
82
+ -110.7421875,
83
+ -19.599609375,
84
+ -173.3203125
85
+ ],
86
+ "mean": [
87
+ -26.84657096862793,
88
+ 4.594353199005127,
89
+ -23.747364044189453,
90
+ 74.75859832763672,
91
+ -111.30194854736328
92
+ ],
93
+ "std": [
94
+ 27.627857208251953,
95
+ 40.672019958496094,
96
+ 48.966896057128906,
97
+ 20.728727340698242,
98
+ 46.57689666748047
99
+ ],
100
+ "q01": [
101
+ -79.453125,
102
+ -102.744140625,
103
+ -109.6875,
104
+ 17.9296875,
105
+ -172.177734375
106
+ ],
107
+ "q99": [
108
+ 24.08203125,
109
+ 72.685546875,
110
+ 83.84765625,
111
+ 101.77734375,
112
+ 10.72265625
113
+ ]
114
+ },
115
+ "gripper": {
116
+ "max": [
117
+ 92.63671875
118
+ ],
119
+ "min": [
120
+ -3.076171875
121
+ ],
122
+ "mean": [
123
+ 12.46182632446289
124
+ ],
125
+ "std": [
126
+ 14.029601097106934
127
+ ],
128
+ "q01": [
129
+ -0.52734375
130
+ ],
131
+ "q99": [
132
+ 52.91015625
133
+ ]
134
+ }
135
+ }
136
+ },
137
+ "modalities": {
138
+ "video": {
139
+ "front": {
140
+ "resolution": [
141
+ 640,
142
+ 480
143
+ ],
144
+ "channels": 3,
145
+ "fps": 30.0
146
+ },
147
+ "laptop": {
148
+ "resolution": [
149
+ 640,
150
+ 480
151
+ ],
152
+ "channels": 3,
153
+ "fps": 30.0
154
+ }
155
+ },
156
+ "state": {
157
+ "single_arm": {
158
+ "absolute": true,
159
+ "rotation_type": null,
160
+ "shape": [
161
+ 5
162
+ ],
163
+ "continuous": true
164
+ },
165
+ "gripper": {
166
+ "absolute": true,
167
+ "rotation_type": null,
168
+ "shape": [
169
+ 1
170
+ ],
171
+ "continuous": true
172
+ }
173
+ },
174
+ "action": {
175
+ "single_arm": {
176
+ "absolute": true,
177
+ "rotation_type": null,
178
+ "shape": [
179
+ 5
180
+ ],
181
+ "continuous": true
182
+ },
183
+ "gripper": {
184
+ "absolute": true,
185
+ "rotation_type": null,
186
+ "shape": [
187
+ 1
188
+ ],
189
+ "continuous": true
190
+ }
191
+ }
192
+ },
193
+ "embodiment_tag": "new_embodiment"
194
+ }
195
+ }
checkpoint-5000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0e65895a7b6c052e58b679bdd68bff54a68694a81f38e23d3e03a92059b3600b
3
+ size 4999367032
checkpoint-5000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:942142544a489d10a803779874cba9a1ec18d2cc3ccabfe9f5c2879f5eedfe04
3
+ size 2586705312
checkpoint-5000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-5000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:135deb8770b61632d6a636310cc1589a33722ffdda34d98861c3ee95512da7ac
3
+ size 4147424698
checkpoint-5000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e89704cf03f557538e476cee66579fbaadf78bed5388d837bfa424566e8e0755
3
+ size 14244
checkpoint-5000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f3860515cdc95c999d5fee0e8633bc5123ccc0bc4f65d625f0a33425450fa0f7
3
+ size 1064
checkpoint-5000/trainer_state.json ADDED
@@ -0,0 +1,3534 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 1.9912385503783354,
6
+ "eval_steps": 500,
7
+ "global_step": 5000,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.00398247710075667,
14
+ "grad_norm": 10.328946113586426,
15
+ "learning_rate": 1.8e-06,
16
+ "loss": 0.9415,
17
+ "step": 10
18
+ },
19
+ {
20
+ "epoch": 0.00796495420151334,
21
+ "grad_norm": 6.881450176239014,
22
+ "learning_rate": 3.8e-06,
23
+ "loss": 0.8639,
24
+ "step": 20
25
+ },
26
+ {
27
+ "epoch": 0.011947431302270013,
28
+ "grad_norm": 2.251582384109497,
29
+ "learning_rate": 5.8e-06,
30
+ "loss": 0.4771,
31
+ "step": 30
32
+ },
33
+ {
34
+ "epoch": 0.01592990840302668,
35
+ "grad_norm": 1.6325163841247559,
36
+ "learning_rate": 7.8e-06,
37
+ "loss": 0.2953,
38
+ "step": 40
39
+ },
40
+ {
41
+ "epoch": 0.019912385503783353,
42
+ "grad_norm": 0.9643872976303101,
43
+ "learning_rate": 9.800000000000001e-06,
44
+ "loss": 0.2403,
45
+ "step": 50
46
+ },
47
+ {
48
+ "epoch": 0.023894862604540025,
49
+ "grad_norm": 1.4014142751693726,
50
+ "learning_rate": 1.18e-05,
51
+ "loss": 0.2213,
52
+ "step": 60
53
+ },
54
+ {
55
+ "epoch": 0.027877339705296694,
56
+ "grad_norm": 1.3301987648010254,
57
+ "learning_rate": 1.3800000000000002e-05,
58
+ "loss": 0.2048,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.03185981680605336,
63
+ "grad_norm": 1.5023683309555054,
64
+ "learning_rate": 1.58e-05,
65
+ "loss": 0.1658,
66
+ "step": 80
67
+ },
68
+ {
69
+ "epoch": 0.035842293906810034,
70
+ "grad_norm": 1.7575840950012207,
71
+ "learning_rate": 1.78e-05,
72
+ "loss": 0.146,
73
+ "step": 90
74
+ },
75
+ {
76
+ "epoch": 0.039824771007566706,
77
+ "grad_norm": 1.4768755435943604,
78
+ "learning_rate": 1.9800000000000004e-05,
79
+ "loss": 0.154,
80
+ "step": 100
81
+ },
82
+ {
83
+ "epoch": 0.04380724810832338,
84
+ "grad_norm": 1.43697988986969,
85
+ "learning_rate": 2.18e-05,
86
+ "loss": 0.1419,
87
+ "step": 110
88
+ },
89
+ {
90
+ "epoch": 0.04778972520908005,
91
+ "grad_norm": 1.6363322734832764,
92
+ "learning_rate": 2.38e-05,
93
+ "loss": 0.1243,
94
+ "step": 120
95
+ },
96
+ {
97
+ "epoch": 0.051772202309836715,
98
+ "grad_norm": 1.29106605052948,
99
+ "learning_rate": 2.58e-05,
100
+ "loss": 0.1273,
101
+ "step": 130
102
+ },
103
+ {
104
+ "epoch": 0.05575467941059339,
105
+ "grad_norm": 0.849351704120636,
106
+ "learning_rate": 2.7800000000000005e-05,
107
+ "loss": 0.1199,
108
+ "step": 140
109
+ },
110
+ {
111
+ "epoch": 0.05973715651135006,
112
+ "grad_norm": 1.442665934562683,
113
+ "learning_rate": 2.98e-05,
114
+ "loss": 0.1068,
115
+ "step": 150
116
+ },
117
+ {
118
+ "epoch": 0.06371963361210672,
119
+ "grad_norm": 1.277969241142273,
120
+ "learning_rate": 3.18e-05,
121
+ "loss": 0.11,
122
+ "step": 160
123
+ },
124
+ {
125
+ "epoch": 0.0677021107128634,
126
+ "grad_norm": 1.636638879776001,
127
+ "learning_rate": 3.38e-05,
128
+ "loss": 0.1001,
129
+ "step": 170
130
+ },
131
+ {
132
+ "epoch": 0.07168458781362007,
133
+ "grad_norm": 1.0019601583480835,
134
+ "learning_rate": 3.58e-05,
135
+ "loss": 0.1126,
136
+ "step": 180
137
+ },
138
+ {
139
+ "epoch": 0.07566706491437675,
140
+ "grad_norm": 0.9693198204040527,
141
+ "learning_rate": 3.7800000000000004e-05,
142
+ "loss": 0.0867,
143
+ "step": 190
144
+ },
145
+ {
146
+ "epoch": 0.07964954201513341,
147
+ "grad_norm": 0.7893789410591125,
148
+ "learning_rate": 3.9800000000000005e-05,
149
+ "loss": 0.1002,
150
+ "step": 200
151
+ },
152
+ {
153
+ "epoch": 0.08363201911589008,
154
+ "grad_norm": 1.736881136894226,
155
+ "learning_rate": 4.18e-05,
156
+ "loss": 0.1015,
157
+ "step": 210
158
+ },
159
+ {
160
+ "epoch": 0.08761449621664676,
161
+ "grad_norm": 0.7016027569770813,
162
+ "learning_rate": 4.38e-05,
163
+ "loss": 0.0895,
164
+ "step": 220
165
+ },
166
+ {
167
+ "epoch": 0.09159697331740342,
168
+ "grad_norm": 1.6043905019760132,
169
+ "learning_rate": 4.58e-05,
170
+ "loss": 0.0857,
171
+ "step": 230
172
+ },
173
+ {
174
+ "epoch": 0.0955794504181601,
175
+ "grad_norm": 1.4260313510894775,
176
+ "learning_rate": 4.78e-05,
177
+ "loss": 0.0818,
178
+ "step": 240
179
+ },
180
+ {
181
+ "epoch": 0.09956192751891677,
182
+ "grad_norm": 0.9992408156394958,
183
+ "learning_rate": 4.9800000000000004e-05,
184
+ "loss": 0.0763,
185
+ "step": 250
186
+ },
187
+ {
188
+ "epoch": 0.10354440461967343,
189
+ "grad_norm": 0.5259194374084473,
190
+ "learning_rate": 5.1800000000000005e-05,
191
+ "loss": 0.0728,
192
+ "step": 260
193
+ },
194
+ {
195
+ "epoch": 0.10752688172043011,
196
+ "grad_norm": 0.8904138207435608,
197
+ "learning_rate": 5.380000000000001e-05,
198
+ "loss": 0.0741,
199
+ "step": 270
200
+ },
201
+ {
202
+ "epoch": 0.11150935882118677,
203
+ "grad_norm": 0.8944669961929321,
204
+ "learning_rate": 5.580000000000001e-05,
205
+ "loss": 0.0646,
206
+ "step": 280
207
+ },
208
+ {
209
+ "epoch": 0.11549183592194345,
210
+ "grad_norm": 1.0091243982315063,
211
+ "learning_rate": 5.7799999999999995e-05,
212
+ "loss": 0.07,
213
+ "step": 290
214
+ },
215
+ {
216
+ "epoch": 0.11947431302270012,
217
+ "grad_norm": 0.7492197155952454,
218
+ "learning_rate": 5.9800000000000003e-05,
219
+ "loss": 0.0726,
220
+ "step": 300
221
+ },
222
+ {
223
+ "epoch": 0.12345679012345678,
224
+ "grad_norm": 0.9669196605682373,
225
+ "learning_rate": 6.18e-05,
226
+ "loss": 0.068,
227
+ "step": 310
228
+ },
229
+ {
230
+ "epoch": 0.12743926722421345,
231
+ "grad_norm": 0.8658894896507263,
232
+ "learning_rate": 6.38e-05,
233
+ "loss": 0.0605,
234
+ "step": 320
235
+ },
236
+ {
237
+ "epoch": 0.13142174432497014,
238
+ "grad_norm": 0.6594343185424805,
239
+ "learning_rate": 6.58e-05,
240
+ "loss": 0.06,
241
+ "step": 330
242
+ },
243
+ {
244
+ "epoch": 0.1354042214257268,
245
+ "grad_norm": 0.9863249659538269,
246
+ "learning_rate": 6.780000000000001e-05,
247
+ "loss": 0.0582,
248
+ "step": 340
249
+ },
250
+ {
251
+ "epoch": 0.13938669852648347,
252
+ "grad_norm": 0.8012584447860718,
253
+ "learning_rate": 6.98e-05,
254
+ "loss": 0.0581,
255
+ "step": 350
256
+ },
257
+ {
258
+ "epoch": 0.14336917562724014,
259
+ "grad_norm": 0.8694135546684265,
260
+ "learning_rate": 7.18e-05,
261
+ "loss": 0.06,
262
+ "step": 360
263
+ },
264
+ {
265
+ "epoch": 0.1473516527279968,
266
+ "grad_norm": 0.8341110944747925,
267
+ "learning_rate": 7.38e-05,
268
+ "loss": 0.0551,
269
+ "step": 370
270
+ },
271
+ {
272
+ "epoch": 0.1513341298287535,
273
+ "grad_norm": 0.8546151518821716,
274
+ "learning_rate": 7.58e-05,
275
+ "loss": 0.0504,
276
+ "step": 380
277
+ },
278
+ {
279
+ "epoch": 0.15531660692951016,
280
+ "grad_norm": 0.8565087914466858,
281
+ "learning_rate": 7.780000000000001e-05,
282
+ "loss": 0.0544,
283
+ "step": 390
284
+ },
285
+ {
286
+ "epoch": 0.15929908403026682,
287
+ "grad_norm": 0.9835138320922852,
288
+ "learning_rate": 7.98e-05,
289
+ "loss": 0.0528,
290
+ "step": 400
291
+ },
292
+ {
293
+ "epoch": 0.1632815611310235,
294
+ "grad_norm": 0.9852800965309143,
295
+ "learning_rate": 8.18e-05,
296
+ "loss": 0.0556,
297
+ "step": 410
298
+ },
299
+ {
300
+ "epoch": 0.16726403823178015,
301
+ "grad_norm": 0.9271975755691528,
302
+ "learning_rate": 8.38e-05,
303
+ "loss": 0.053,
304
+ "step": 420
305
+ },
306
+ {
307
+ "epoch": 0.17124651533253685,
308
+ "grad_norm": 0.843838632106781,
309
+ "learning_rate": 8.58e-05,
310
+ "loss": 0.0536,
311
+ "step": 430
312
+ },
313
+ {
314
+ "epoch": 0.1752289924332935,
315
+ "grad_norm": 0.6337900161743164,
316
+ "learning_rate": 8.78e-05,
317
+ "loss": 0.0501,
318
+ "step": 440
319
+ },
320
+ {
321
+ "epoch": 0.17921146953405018,
322
+ "grad_norm": 0.8339642882347107,
323
+ "learning_rate": 8.98e-05,
324
+ "loss": 0.0535,
325
+ "step": 450
326
+ },
327
+ {
328
+ "epoch": 0.18319394663480684,
329
+ "grad_norm": 0.8911416530609131,
330
+ "learning_rate": 9.180000000000001e-05,
331
+ "loss": 0.0599,
332
+ "step": 460
333
+ },
334
+ {
335
+ "epoch": 0.1871764237355635,
336
+ "grad_norm": 0.7793362736701965,
337
+ "learning_rate": 9.38e-05,
338
+ "loss": 0.0585,
339
+ "step": 470
340
+ },
341
+ {
342
+ "epoch": 0.1911589008363202,
343
+ "grad_norm": 1.1783093214035034,
344
+ "learning_rate": 9.58e-05,
345
+ "loss": 0.0509,
346
+ "step": 480
347
+ },
348
+ {
349
+ "epoch": 0.19514137793707687,
350
+ "grad_norm": 0.9017106294631958,
351
+ "learning_rate": 9.78e-05,
352
+ "loss": 0.0558,
353
+ "step": 490
354
+ },
355
+ {
356
+ "epoch": 0.19912385503783353,
357
+ "grad_norm": 1.0761146545410156,
358
+ "learning_rate": 9.98e-05,
359
+ "loss": 0.0519,
360
+ "step": 500
361
+ },
362
+ {
363
+ "epoch": 0.2031063321385902,
364
+ "grad_norm": 0.5883626937866211,
365
+ "learning_rate": 9.9999778549206e-05,
366
+ "loss": 0.0527,
367
+ "step": 510
368
+ },
369
+ {
370
+ "epoch": 0.20708880923934686,
371
+ "grad_norm": 0.6159175634384155,
372
+ "learning_rate": 9.999901304280685e-05,
373
+ "loss": 0.0488,
374
+ "step": 520
375
+ },
376
+ {
377
+ "epoch": 0.21107128634010355,
378
+ "grad_norm": 0.6942933797836304,
379
+ "learning_rate": 9.999770075521164e-05,
380
+ "loss": 0.0524,
381
+ "step": 530
382
+ },
383
+ {
384
+ "epoch": 0.21505376344086022,
385
+ "grad_norm": 1.0295978784561157,
386
+ "learning_rate": 9.99958417007713e-05,
387
+ "loss": 0.049,
388
+ "step": 540
389
+ },
390
+ {
391
+ "epoch": 0.21903624054161688,
392
+ "grad_norm": 0.7261350154876709,
393
+ "learning_rate": 9.999343589981615e-05,
394
+ "loss": 0.0467,
395
+ "step": 550
396
+ },
397
+ {
398
+ "epoch": 0.22301871764237355,
399
+ "grad_norm": 0.7238365411758423,
400
+ "learning_rate": 9.999048337865568e-05,
401
+ "loss": 0.0477,
402
+ "step": 560
403
+ },
404
+ {
405
+ "epoch": 0.2270011947431302,
406
+ "grad_norm": 1.1104141473770142,
407
+ "learning_rate": 9.998698416957815e-05,
408
+ "loss": 0.0519,
409
+ "step": 570
410
+ },
411
+ {
412
+ "epoch": 0.2309836718438869,
413
+ "grad_norm": 1.2086730003356934,
414
+ "learning_rate": 9.998293831085037e-05,
415
+ "loss": 0.0524,
416
+ "step": 580
417
+ },
418
+ {
419
+ "epoch": 0.23496614894464357,
420
+ "grad_norm": 0.5849373936653137,
421
+ "learning_rate": 9.997834584671719e-05,
422
+ "loss": 0.0514,
423
+ "step": 590
424
+ },
425
+ {
426
+ "epoch": 0.23894862604540024,
427
+ "grad_norm": 0.9337085485458374,
428
+ "learning_rate": 9.997320682740107e-05,
429
+ "loss": 0.0565,
430
+ "step": 600
431
+ },
432
+ {
433
+ "epoch": 0.2429311031461569,
434
+ "grad_norm": 0.7592830061912537,
435
+ "learning_rate": 9.996752130910149e-05,
436
+ "loss": 0.0521,
437
+ "step": 610
438
+ },
439
+ {
440
+ "epoch": 0.24691358024691357,
441
+ "grad_norm": 0.8495620489120483,
442
+ "learning_rate": 9.99612893539944e-05,
443
+ "loss": 0.0501,
444
+ "step": 620
445
+ },
446
+ {
447
+ "epoch": 0.25089605734767023,
448
+ "grad_norm": 0.44615253806114197,
449
+ "learning_rate": 9.995451103023144e-05,
450
+ "loss": 0.0428,
451
+ "step": 630
452
+ },
453
+ {
454
+ "epoch": 0.2548785344484269,
455
+ "grad_norm": 1.046374797821045,
456
+ "learning_rate": 9.994718641193928e-05,
457
+ "loss": 0.0459,
458
+ "step": 640
459
+ },
460
+ {
461
+ "epoch": 0.2588610115491836,
462
+ "grad_norm": 0.7210455536842346,
463
+ "learning_rate": 9.993931557921874e-05,
464
+ "loss": 0.0508,
465
+ "step": 650
466
+ },
467
+ {
468
+ "epoch": 0.2628434886499403,
469
+ "grad_norm": 0.7506885528564453,
470
+ "learning_rate": 9.993089861814402e-05,
471
+ "loss": 0.0499,
472
+ "step": 660
473
+ },
474
+ {
475
+ "epoch": 0.26682596575069695,
476
+ "grad_norm": 0.6320833563804626,
477
+ "learning_rate": 9.992193562076166e-05,
478
+ "loss": 0.0579,
479
+ "step": 670
480
+ },
481
+ {
482
+ "epoch": 0.2708084428514536,
483
+ "grad_norm": 0.48116907477378845,
484
+ "learning_rate": 9.991242668508954e-05,
485
+ "loss": 0.0533,
486
+ "step": 680
487
+ },
488
+ {
489
+ "epoch": 0.2747909199522103,
490
+ "grad_norm": 0.5594539046287537,
491
+ "learning_rate": 9.990237191511587e-05,
492
+ "loss": 0.0479,
493
+ "step": 690
494
+ },
495
+ {
496
+ "epoch": 0.27877339705296694,
497
+ "grad_norm": 0.6126905083656311,
498
+ "learning_rate": 9.989177142079802e-05,
499
+ "loss": 0.0449,
500
+ "step": 700
501
+ },
502
+ {
503
+ "epoch": 0.2827558741537236,
504
+ "grad_norm": 0.8031131625175476,
505
+ "learning_rate": 9.988062531806126e-05,
506
+ "loss": 0.0485,
507
+ "step": 710
508
+ },
509
+ {
510
+ "epoch": 0.2867383512544803,
511
+ "grad_norm": 0.4269757568836212,
512
+ "learning_rate": 9.986893372879762e-05,
513
+ "loss": 0.0463,
514
+ "step": 720
515
+ },
516
+ {
517
+ "epoch": 0.29072082835523694,
518
+ "grad_norm": 0.8300117254257202,
519
+ "learning_rate": 9.985669678086443e-05,
520
+ "loss": 0.0505,
521
+ "step": 730
522
+ },
523
+ {
524
+ "epoch": 0.2947033054559936,
525
+ "grad_norm": 0.6425809264183044,
526
+ "learning_rate": 9.984391460808298e-05,
527
+ "loss": 0.0407,
528
+ "step": 740
529
+ },
530
+ {
531
+ "epoch": 0.2986857825567503,
532
+ "grad_norm": 0.4019189774990082,
533
+ "learning_rate": 9.983058735023709e-05,
534
+ "loss": 0.044,
535
+ "step": 750
536
+ },
537
+ {
538
+ "epoch": 0.302668259657507,
539
+ "grad_norm": 1.2118290662765503,
540
+ "learning_rate": 9.98167151530715e-05,
541
+ "loss": 0.0435,
542
+ "step": 760
543
+ },
544
+ {
545
+ "epoch": 0.30665073675826365,
546
+ "grad_norm": 0.4535205662250519,
547
+ "learning_rate": 9.980229816829034e-05,
548
+ "loss": 0.043,
549
+ "step": 770
550
+ },
551
+ {
552
+ "epoch": 0.3106332138590203,
553
+ "grad_norm": 0.7236475348472595,
554
+ "learning_rate": 9.978733655355544e-05,
555
+ "loss": 0.0479,
556
+ "step": 780
557
+ },
558
+ {
559
+ "epoch": 0.314615690959777,
560
+ "grad_norm": 1.045904278755188,
561
+ "learning_rate": 9.977183047248464e-05,
562
+ "loss": 0.0458,
563
+ "step": 790
564
+ },
565
+ {
566
+ "epoch": 0.31859816806053365,
567
+ "grad_norm": 0.8453964591026306,
568
+ "learning_rate": 9.975578009464992e-05,
569
+ "loss": 0.0489,
570
+ "step": 800
571
+ },
572
+ {
573
+ "epoch": 0.3225806451612903,
574
+ "grad_norm": 0.8722575902938843,
575
+ "learning_rate": 9.97391855955757e-05,
576
+ "loss": 0.0496,
577
+ "step": 810
578
+ },
579
+ {
580
+ "epoch": 0.326563122262047,
581
+ "grad_norm": 0.5823293328285217,
582
+ "learning_rate": 9.972204715673669e-05,
583
+ "loss": 0.0444,
584
+ "step": 820
585
+ },
586
+ {
587
+ "epoch": 0.33054559936280364,
588
+ "grad_norm": 0.426830917596817,
589
+ "learning_rate": 9.970436496555617e-05,
590
+ "loss": 0.0429,
591
+ "step": 830
592
+ },
593
+ {
594
+ "epoch": 0.3345280764635603,
595
+ "grad_norm": 0.6528297066688538,
596
+ "learning_rate": 9.968613921540373e-05,
597
+ "loss": 0.0462,
598
+ "step": 840
599
+ },
600
+ {
601
+ "epoch": 0.33851055356431703,
602
+ "grad_norm": 0.8501228094100952,
603
+ "learning_rate": 9.966737010559326e-05,
604
+ "loss": 0.0483,
605
+ "step": 850
606
+ },
607
+ {
608
+ "epoch": 0.3424930306650737,
609
+ "grad_norm": 0.6673285365104675,
610
+ "learning_rate": 9.964805784138072e-05,
611
+ "loss": 0.0483,
612
+ "step": 860
613
+ },
614
+ {
615
+ "epoch": 0.34647550776583036,
616
+ "grad_norm": 0.6113673448562622,
617
+ "learning_rate": 9.962820263396195e-05,
618
+ "loss": 0.0439,
619
+ "step": 870
620
+ },
621
+ {
622
+ "epoch": 0.350457984866587,
623
+ "grad_norm": 0.8236457705497742,
624
+ "learning_rate": 9.960780470047033e-05,
625
+ "loss": 0.0425,
626
+ "step": 880
627
+ },
628
+ {
629
+ "epoch": 0.3544404619673437,
630
+ "grad_norm": 0.4798528552055359,
631
+ "learning_rate": 9.958686426397437e-05,
632
+ "loss": 0.0397,
633
+ "step": 890
634
+ },
635
+ {
636
+ "epoch": 0.35842293906810035,
637
+ "grad_norm": 0.5209172964096069,
638
+ "learning_rate": 9.956538155347534e-05,
639
+ "loss": 0.0496,
640
+ "step": 900
641
+ },
642
+ {
643
+ "epoch": 0.362405416168857,
644
+ "grad_norm": 0.549393892288208,
645
+ "learning_rate": 9.95433568039047e-05,
646
+ "loss": 0.0388,
647
+ "step": 910
648
+ },
649
+ {
650
+ "epoch": 0.3663878932696137,
651
+ "grad_norm": 0.6078306436538696,
652
+ "learning_rate": 9.952079025612162e-05,
653
+ "loss": 0.0467,
654
+ "step": 920
655
+ },
656
+ {
657
+ "epoch": 0.37037037037037035,
658
+ "grad_norm": 0.47643283009529114,
659
+ "learning_rate": 9.949768215691022e-05,
660
+ "loss": 0.0392,
661
+ "step": 930
662
+ },
663
+ {
664
+ "epoch": 0.374352847471127,
665
+ "grad_norm": 0.5230958461761475,
666
+ "learning_rate": 9.9474032758977e-05,
667
+ "loss": 0.0414,
668
+ "step": 940
669
+ },
670
+ {
671
+ "epoch": 0.37833532457188374,
672
+ "grad_norm": 0.5619562864303589,
673
+ "learning_rate": 9.944984232094794e-05,
674
+ "loss": 0.0448,
675
+ "step": 950
676
+ },
677
+ {
678
+ "epoch": 0.3823178016726404,
679
+ "grad_norm": 0.7934162616729736,
680
+ "learning_rate": 9.942511110736584e-05,
681
+ "loss": 0.0389,
682
+ "step": 960
683
+ },
684
+ {
685
+ "epoch": 0.38630027877339707,
686
+ "grad_norm": 0.7610427737236023,
687
+ "learning_rate": 9.939983938868726e-05,
688
+ "loss": 0.0383,
689
+ "step": 970
690
+ },
691
+ {
692
+ "epoch": 0.39028275587415373,
693
+ "grad_norm": 0.6311072111129761,
694
+ "learning_rate": 9.93740274412797e-05,
695
+ "loss": 0.0449,
696
+ "step": 980
697
+ },
698
+ {
699
+ "epoch": 0.3942652329749104,
700
+ "grad_norm": 0.3925153613090515,
701
+ "learning_rate": 9.934767554741846e-05,
702
+ "loss": 0.043,
703
+ "step": 990
704
+ },
705
+ {
706
+ "epoch": 0.39824771007566706,
707
+ "grad_norm": 0.4677866995334625,
708
+ "learning_rate": 9.932078399528361e-05,
709
+ "loss": 0.0414,
710
+ "step": 1000
711
+ },
712
+ {
713
+ "epoch": 0.4022301871764237,
714
+ "grad_norm": 0.38546767830848694,
715
+ "learning_rate": 9.929335307895689e-05,
716
+ "loss": 0.0342,
717
+ "step": 1010
718
+ },
719
+ {
720
+ "epoch": 0.4062126642771804,
721
+ "grad_norm": 0.5453174710273743,
722
+ "learning_rate": 9.926538309841839e-05,
723
+ "loss": 0.0421,
724
+ "step": 1020
725
+ },
726
+ {
727
+ "epoch": 0.41019514137793706,
728
+ "grad_norm": 0.34743279218673706,
729
+ "learning_rate": 9.923687435954334e-05,
730
+ "loss": 0.0387,
731
+ "step": 1030
732
+ },
733
+ {
734
+ "epoch": 0.4141776184786937,
735
+ "grad_norm": 0.554025411605835,
736
+ "learning_rate": 9.920782717409873e-05,
737
+ "loss": 0.0469,
738
+ "step": 1040
739
+ },
740
+ {
741
+ "epoch": 0.41816009557945044,
742
+ "grad_norm": 0.6354169249534607,
743
+ "learning_rate": 9.917824185973994e-05,
744
+ "loss": 0.0411,
745
+ "step": 1050
746
+ },
747
+ {
748
+ "epoch": 0.4221425726802071,
749
+ "grad_norm": 0.46389538049697876,
750
+ "learning_rate": 9.914811874000723e-05,
751
+ "loss": 0.0447,
752
+ "step": 1060
753
+ },
754
+ {
755
+ "epoch": 0.42612504978096377,
756
+ "grad_norm": 0.8935972452163696,
757
+ "learning_rate": 9.911745814432218e-05,
758
+ "loss": 0.0454,
759
+ "step": 1070
760
+ },
761
+ {
762
+ "epoch": 0.43010752688172044,
763
+ "grad_norm": 1.0306254625320435,
764
+ "learning_rate": 9.90862604079842e-05,
765
+ "loss": 0.0438,
766
+ "step": 1080
767
+ },
768
+ {
769
+ "epoch": 0.4340900039824771,
770
+ "grad_norm": 0.5918930768966675,
771
+ "learning_rate": 9.90545258721667e-05,
772
+ "loss": 0.0389,
773
+ "step": 1090
774
+ },
775
+ {
776
+ "epoch": 0.43807248108323377,
777
+ "grad_norm": 0.5646054148674011,
778
+ "learning_rate": 9.90222548839135e-05,
779
+ "loss": 0.0395,
780
+ "step": 1100
781
+ },
782
+ {
783
+ "epoch": 0.44205495818399043,
784
+ "grad_norm": 0.5279185175895691,
785
+ "learning_rate": 9.898944779613495e-05,
786
+ "loss": 0.0438,
787
+ "step": 1110
788
+ },
789
+ {
790
+ "epoch": 0.4460374352847471,
791
+ "grad_norm": 0.41257861256599426,
792
+ "learning_rate": 9.89561049676041e-05,
793
+ "loss": 0.0397,
794
+ "step": 1120
795
+ },
796
+ {
797
+ "epoch": 0.45001991238550376,
798
+ "grad_norm": 0.36252203583717346,
799
+ "learning_rate": 9.89222267629528e-05,
800
+ "loss": 0.043,
801
+ "step": 1130
802
+ },
803
+ {
804
+ "epoch": 0.4540023894862604,
805
+ "grad_norm": 0.8424637913703918,
806
+ "learning_rate": 9.888781355266763e-05,
807
+ "loss": 0.0433,
808
+ "step": 1140
809
+ },
810
+ {
811
+ "epoch": 0.45798486658701715,
812
+ "grad_norm": 0.5604090690612793,
813
+ "learning_rate": 9.885286571308598e-05,
814
+ "loss": 0.0452,
815
+ "step": 1150
816
+ },
817
+ {
818
+ "epoch": 0.4619673436877738,
819
+ "grad_norm": 0.7178535461425781,
820
+ "learning_rate": 9.881738362639182e-05,
821
+ "loss": 0.0401,
822
+ "step": 1160
823
+ },
824
+ {
825
+ "epoch": 0.4659498207885305,
826
+ "grad_norm": 0.7291957139968872,
827
+ "learning_rate": 9.878136768061154e-05,
828
+ "loss": 0.0368,
829
+ "step": 1170
830
+ },
831
+ {
832
+ "epoch": 0.46993229788928714,
833
+ "grad_norm": 0.48816534876823425,
834
+ "learning_rate": 9.874481826960979e-05,
835
+ "loss": 0.039,
836
+ "step": 1180
837
+ },
838
+ {
839
+ "epoch": 0.4739147749900438,
840
+ "grad_norm": 0.813832700252533,
841
+ "learning_rate": 9.870773579308503e-05,
842
+ "loss": 0.0417,
843
+ "step": 1190
844
+ },
845
+ {
846
+ "epoch": 0.4778972520908005,
847
+ "grad_norm": 0.5479532480239868,
848
+ "learning_rate": 9.867012065656533e-05,
849
+ "loss": 0.0428,
850
+ "step": 1200
851
+ },
852
+ {
853
+ "epoch": 0.48187972919155714,
854
+ "grad_norm": 0.7760400772094727,
855
+ "learning_rate": 9.863197327140376e-05,
856
+ "loss": 0.0454,
857
+ "step": 1210
858
+ },
859
+ {
860
+ "epoch": 0.4858622062923138,
861
+ "grad_norm": 0.6200589537620544,
862
+ "learning_rate": 9.859329405477403e-05,
863
+ "loss": 0.0377,
864
+ "step": 1220
865
+ },
866
+ {
867
+ "epoch": 0.48984468339307047,
868
+ "grad_norm": 0.7709351778030396,
869
+ "learning_rate": 9.855408342966585e-05,
870
+ "loss": 0.045,
871
+ "step": 1230
872
+ },
873
+ {
874
+ "epoch": 0.49382716049382713,
875
+ "grad_norm": 0.44105854630470276,
876
+ "learning_rate": 9.851434182488033e-05,
877
+ "loss": 0.0374,
878
+ "step": 1240
879
+ },
880
+ {
881
+ "epoch": 0.49780963759458385,
882
+ "grad_norm": 0.4224795401096344,
883
+ "learning_rate": 9.84740696750253e-05,
884
+ "loss": 0.0403,
885
+ "step": 1250
886
+ },
887
+ {
888
+ "epoch": 0.5017921146953405,
889
+ "grad_norm": 0.49829116463661194,
890
+ "learning_rate": 9.843326742051055e-05,
891
+ "loss": 0.0427,
892
+ "step": 1260
893
+ },
894
+ {
895
+ "epoch": 0.5057745917960972,
896
+ "grad_norm": 0.42388513684272766,
897
+ "learning_rate": 9.839193550754297e-05,
898
+ "loss": 0.0407,
899
+ "step": 1270
900
+ },
901
+ {
902
+ "epoch": 0.5097570688968538,
903
+ "grad_norm": 0.8282580375671387,
904
+ "learning_rate": 9.835007438812177e-05,
905
+ "loss": 0.0422,
906
+ "step": 1280
907
+ },
908
+ {
909
+ "epoch": 0.5137395459976105,
910
+ "grad_norm": 0.7779645323753357,
911
+ "learning_rate": 9.830768452003341e-05,
912
+ "loss": 0.044,
913
+ "step": 1290
914
+ },
915
+ {
916
+ "epoch": 0.5177220230983672,
917
+ "grad_norm": 0.6992411017417908,
918
+ "learning_rate": 9.826476636684671e-05,
919
+ "loss": 0.0422,
920
+ "step": 1300
921
+ },
922
+ {
923
+ "epoch": 0.5217045001991238,
924
+ "grad_norm": 0.5714141130447388,
925
+ "learning_rate": 9.822132039790773e-05,
926
+ "loss": 0.0394,
927
+ "step": 1310
928
+ },
929
+ {
930
+ "epoch": 0.5256869772998806,
931
+ "grad_norm": 0.8175871968269348,
932
+ "learning_rate": 9.817734708833461e-05,
933
+ "loss": 0.0498,
934
+ "step": 1320
935
+ },
936
+ {
937
+ "epoch": 0.5296694544006372,
938
+ "grad_norm": 0.6209958791732788,
939
+ "learning_rate": 9.813284691901243e-05,
940
+ "loss": 0.04,
941
+ "step": 1330
942
+ },
943
+ {
944
+ "epoch": 0.5336519315013939,
945
+ "grad_norm": 0.39030686020851135,
946
+ "learning_rate": 9.808782037658792e-05,
947
+ "loss": 0.0414,
948
+ "step": 1340
949
+ },
950
+ {
951
+ "epoch": 0.5376344086021505,
952
+ "grad_norm": 0.5858843922615051,
953
+ "learning_rate": 9.804226795346411e-05,
954
+ "loss": 0.0384,
955
+ "step": 1350
956
+ },
957
+ {
958
+ "epoch": 0.5416168857029072,
959
+ "grad_norm": 0.812506914138794,
960
+ "learning_rate": 9.799619014779503e-05,
961
+ "loss": 0.0428,
962
+ "step": 1360
963
+ },
964
+ {
965
+ "epoch": 0.5455993628036638,
966
+ "grad_norm": 0.5438332557678223,
967
+ "learning_rate": 9.794958746348013e-05,
968
+ "loss": 0.0344,
969
+ "step": 1370
970
+ },
971
+ {
972
+ "epoch": 0.5495818399044206,
973
+ "grad_norm": 0.4357604384422302,
974
+ "learning_rate": 9.790246041015896e-05,
975
+ "loss": 0.0429,
976
+ "step": 1380
977
+ },
978
+ {
979
+ "epoch": 0.5535643170051773,
980
+ "grad_norm": 0.6829274892807007,
981
+ "learning_rate": 9.785480950320538e-05,
982
+ "loss": 0.0352,
983
+ "step": 1390
984
+ },
985
+ {
986
+ "epoch": 0.5575467941059339,
987
+ "grad_norm": 0.46741291880607605,
988
+ "learning_rate": 9.78066352637221e-05,
989
+ "loss": 0.0387,
990
+ "step": 1400
991
+ },
992
+ {
993
+ "epoch": 0.5615292712066906,
994
+ "grad_norm": 0.8722621202468872,
995
+ "learning_rate": 9.775793821853488e-05,
996
+ "loss": 0.0364,
997
+ "step": 1410
998
+ },
999
+ {
1000
+ "epoch": 0.5655117483074472,
1001
+ "grad_norm": 0.6563253402709961,
1002
+ "learning_rate": 9.77087189001868e-05,
1003
+ "loss": 0.0409,
1004
+ "step": 1420
1005
+ },
1006
+ {
1007
+ "epoch": 0.5694942254082039,
1008
+ "grad_norm": 0.6959318518638611,
1009
+ "learning_rate": 9.765897784693243e-05,
1010
+ "loss": 0.0349,
1011
+ "step": 1430
1012
+ },
1013
+ {
1014
+ "epoch": 0.5734767025089605,
1015
+ "grad_norm": 0.34205693006515503,
1016
+ "learning_rate": 9.760871560273197e-05,
1017
+ "loss": 0.036,
1018
+ "step": 1440
1019
+ },
1020
+ {
1021
+ "epoch": 0.5774591796097173,
1022
+ "grad_norm": 0.7076167464256287,
1023
+ "learning_rate": 9.755793271724526e-05,
1024
+ "loss": 0.0413,
1025
+ "step": 1450
1026
+ },
1027
+ {
1028
+ "epoch": 0.5814416567104739,
1029
+ "grad_norm": 0.6335232257843018,
1030
+ "learning_rate": 9.750662974582584e-05,
1031
+ "loss": 0.0392,
1032
+ "step": 1460
1033
+ },
1034
+ {
1035
+ "epoch": 0.5854241338112306,
1036
+ "grad_norm": 0.7028225660324097,
1037
+ "learning_rate": 9.745480724951473e-05,
1038
+ "loss": 0.0391,
1039
+ "step": 1470
1040
+ },
1041
+ {
1042
+ "epoch": 0.5894066109119872,
1043
+ "grad_norm": 0.5494594573974609,
1044
+ "learning_rate": 9.740246579503447e-05,
1045
+ "loss": 0.0469,
1046
+ "step": 1480
1047
+ },
1048
+ {
1049
+ "epoch": 0.5933890880127439,
1050
+ "grad_norm": 0.3793744146823883,
1051
+ "learning_rate": 9.734960595478284e-05,
1052
+ "loss": 0.0372,
1053
+ "step": 1490
1054
+ },
1055
+ {
1056
+ "epoch": 0.5973715651135006,
1057
+ "grad_norm": 0.6068878769874573,
1058
+ "learning_rate": 9.729622830682657e-05,
1059
+ "loss": 0.0443,
1060
+ "step": 1500
1061
+ },
1062
+ {
1063
+ "epoch": 0.6013540422142573,
1064
+ "grad_norm": 0.6348056197166443,
1065
+ "learning_rate": 9.724233343489504e-05,
1066
+ "loss": 0.0388,
1067
+ "step": 1510
1068
+ },
1069
+ {
1070
+ "epoch": 0.605336519315014,
1071
+ "grad_norm": 0.7185152173042297,
1072
+ "learning_rate": 9.718792192837396e-05,
1073
+ "loss": 0.0394,
1074
+ "step": 1520
1075
+ },
1076
+ {
1077
+ "epoch": 0.6093189964157706,
1078
+ "grad_norm": 0.29940369725227356,
1079
+ "learning_rate": 9.713299438229886e-05,
1080
+ "loss": 0.0373,
1081
+ "step": 1530
1082
+ },
1083
+ {
1084
+ "epoch": 0.6133014735165273,
1085
+ "grad_norm": 0.608589768409729,
1086
+ "learning_rate": 9.707755139734855e-05,
1087
+ "loss": 0.0444,
1088
+ "step": 1540
1089
+ },
1090
+ {
1091
+ "epoch": 0.6172839506172839,
1092
+ "grad_norm": 0.6535037755966187,
1093
+ "learning_rate": 9.702159357983866e-05,
1094
+ "loss": 0.0418,
1095
+ "step": 1550
1096
+ },
1097
+ {
1098
+ "epoch": 0.6212664277180406,
1099
+ "grad_norm": 0.389615923166275,
1100
+ "learning_rate": 9.696512154171492e-05,
1101
+ "loss": 0.0394,
1102
+ "step": 1560
1103
+ },
1104
+ {
1105
+ "epoch": 0.6252489048187972,
1106
+ "grad_norm": 0.5415939092636108,
1107
+ "learning_rate": 9.690813590054645e-05,
1108
+ "loss": 0.0374,
1109
+ "step": 1570
1110
+ },
1111
+ {
1112
+ "epoch": 0.629231381919554,
1113
+ "grad_norm": 0.6598183512687683,
1114
+ "learning_rate": 9.685063727951914e-05,
1115
+ "loss": 0.0364,
1116
+ "step": 1580
1117
+ },
1118
+ {
1119
+ "epoch": 0.6332138590203107,
1120
+ "grad_norm": 0.4346562325954437,
1121
+ "learning_rate": 9.679262630742865e-05,
1122
+ "loss": 0.0368,
1123
+ "step": 1590
1124
+ },
1125
+ {
1126
+ "epoch": 0.6371963361210673,
1127
+ "grad_norm": 0.7525681257247925,
1128
+ "learning_rate": 9.673410361867373e-05,
1129
+ "loss": 0.037,
1130
+ "step": 1600
1131
+ },
1132
+ {
1133
+ "epoch": 0.641178813221824,
1134
+ "grad_norm": 0.4621691405773163,
1135
+ "learning_rate": 9.667506985324909e-05,
1136
+ "loss": 0.0313,
1137
+ "step": 1610
1138
+ },
1139
+ {
1140
+ "epoch": 0.6451612903225806,
1141
+ "grad_norm": 0.5039979219436646,
1142
+ "learning_rate": 9.661552565673855e-05,
1143
+ "loss": 0.0309,
1144
+ "step": 1620
1145
+ },
1146
+ {
1147
+ "epoch": 0.6491437674233373,
1148
+ "grad_norm": 0.29736581444740295,
1149
+ "learning_rate": 9.655547168030789e-05,
1150
+ "loss": 0.0369,
1151
+ "step": 1630
1152
+ },
1153
+ {
1154
+ "epoch": 0.653126244524094,
1155
+ "grad_norm": 0.4153502583503723,
1156
+ "learning_rate": 9.649490858069777e-05,
1157
+ "loss": 0.0349,
1158
+ "step": 1640
1159
+ },
1160
+ {
1161
+ "epoch": 0.6571087216248507,
1162
+ "grad_norm": 0.64329993724823,
1163
+ "learning_rate": 9.643383702021658e-05,
1164
+ "loss": 0.0422,
1165
+ "step": 1650
1166
+ },
1167
+ {
1168
+ "epoch": 0.6610911987256073,
1169
+ "grad_norm": 0.39090660214424133,
1170
+ "learning_rate": 9.637225766673307e-05,
1171
+ "loss": 0.0332,
1172
+ "step": 1660
1173
+ },
1174
+ {
1175
+ "epoch": 0.665073675826364,
1176
+ "grad_norm": 0.4412473440170288,
1177
+ "learning_rate": 9.631017119366922e-05,
1178
+ "loss": 0.0374,
1179
+ "step": 1670
1180
+ },
1181
+ {
1182
+ "epoch": 0.6690561529271206,
1183
+ "grad_norm": 0.28402477502822876,
1184
+ "learning_rate": 9.624757827999273e-05,
1185
+ "loss": 0.0364,
1186
+ "step": 1680
1187
+ },
1188
+ {
1189
+ "epoch": 0.6730386300278773,
1190
+ "grad_norm": 0.5480361580848694,
1191
+ "learning_rate": 9.618447961020971e-05,
1192
+ "loss": 0.0381,
1193
+ "step": 1690
1194
+ },
1195
+ {
1196
+ "epoch": 0.6770211071286341,
1197
+ "grad_norm": 0.6572909355163574,
1198
+ "learning_rate": 9.612087587435707e-05,
1199
+ "loss": 0.037,
1200
+ "step": 1700
1201
+ },
1202
+ {
1203
+ "epoch": 0.6810035842293907,
1204
+ "grad_norm": 0.6667490601539612,
1205
+ "learning_rate": 9.605676776799508e-05,
1206
+ "loss": 0.04,
1207
+ "step": 1710
1208
+ },
1209
+ {
1210
+ "epoch": 0.6849860613301474,
1211
+ "grad_norm": 0.29837551712989807,
1212
+ "learning_rate": 9.599215599219973e-05,
1213
+ "loss": 0.0351,
1214
+ "step": 1720
1215
+ },
1216
+ {
1217
+ "epoch": 0.688968538430904,
1218
+ "grad_norm": 0.6289225220680237,
1219
+ "learning_rate": 9.592704125355505e-05,
1220
+ "loss": 0.0377,
1221
+ "step": 1730
1222
+ },
1223
+ {
1224
+ "epoch": 0.6929510155316607,
1225
+ "grad_norm": 0.4489145576953888,
1226
+ "learning_rate": 9.586142426414538e-05,
1227
+ "loss": 0.0391,
1228
+ "step": 1740
1229
+ },
1230
+ {
1231
+ "epoch": 0.6969334926324173,
1232
+ "grad_norm": 0.5505160093307495,
1233
+ "learning_rate": 9.57953057415476e-05,
1234
+ "loss": 0.0371,
1235
+ "step": 1750
1236
+ },
1237
+ {
1238
+ "epoch": 0.700915969733174,
1239
+ "grad_norm": 0.48338383436203003,
1240
+ "learning_rate": 9.572868640882328e-05,
1241
+ "loss": 0.0399,
1242
+ "step": 1760
1243
+ },
1244
+ {
1245
+ "epoch": 0.7048984468339307,
1246
+ "grad_norm": 0.3950949013233185,
1247
+ "learning_rate": 9.56615669945108e-05,
1248
+ "loss": 0.0411,
1249
+ "step": 1770
1250
+ },
1251
+ {
1252
+ "epoch": 0.7088809239346874,
1253
+ "grad_norm": 0.7186228036880493,
1254
+ "learning_rate": 9.55939482326173e-05,
1255
+ "loss": 0.0358,
1256
+ "step": 1780
1257
+ },
1258
+ {
1259
+ "epoch": 0.7128634010354441,
1260
+ "grad_norm": 0.36578911542892456,
1261
+ "learning_rate": 9.552583086261069e-05,
1262
+ "loss": 0.0402,
1263
+ "step": 1790
1264
+ },
1265
+ {
1266
+ "epoch": 0.7168458781362007,
1267
+ "grad_norm": 0.48990169167518616,
1268
+ "learning_rate": 9.545721562941168e-05,
1269
+ "loss": 0.0353,
1270
+ "step": 1800
1271
+ },
1272
+ {
1273
+ "epoch": 0.7208283552369574,
1274
+ "grad_norm": 0.37718433141708374,
1275
+ "learning_rate": 9.538810328338543e-05,
1276
+ "loss": 0.0385,
1277
+ "step": 1810
1278
+ },
1279
+ {
1280
+ "epoch": 0.724810832337714,
1281
+ "grad_norm": 0.40954500436782837,
1282
+ "learning_rate": 9.531849458033349e-05,
1283
+ "loss": 0.0334,
1284
+ "step": 1820
1285
+ },
1286
+ {
1287
+ "epoch": 0.7287933094384708,
1288
+ "grad_norm": 0.6403422951698303,
1289
+ "learning_rate": 9.524839028148547e-05,
1290
+ "loss": 0.0385,
1291
+ "step": 1830
1292
+ },
1293
+ {
1294
+ "epoch": 0.7327757865392274,
1295
+ "grad_norm": 0.4989095628261566,
1296
+ "learning_rate": 9.517779115349077e-05,
1297
+ "loss": 0.0374,
1298
+ "step": 1840
1299
+ },
1300
+ {
1301
+ "epoch": 0.7367582636399841,
1302
+ "grad_norm": 0.4369076192378998,
1303
+ "learning_rate": 9.510669796841014e-05,
1304
+ "loss": 0.0353,
1305
+ "step": 1850
1306
+ },
1307
+ {
1308
+ "epoch": 0.7407407407407407,
1309
+ "grad_norm": 0.6268195509910583,
1310
+ "learning_rate": 9.503511150370727e-05,
1311
+ "loss": 0.0407,
1312
+ "step": 1860
1313
+ },
1314
+ {
1315
+ "epoch": 0.7447232178414974,
1316
+ "grad_norm": 0.40719932317733765,
1317
+ "learning_rate": 9.496303254224024e-05,
1318
+ "loss": 0.037,
1319
+ "step": 1870
1320
+ },
1321
+ {
1322
+ "epoch": 0.748705694942254,
1323
+ "grad_norm": 0.41097119450569153,
1324
+ "learning_rate": 9.489046187225306e-05,
1325
+ "loss": 0.0338,
1326
+ "step": 1880
1327
+ },
1328
+ {
1329
+ "epoch": 0.7526881720430108,
1330
+ "grad_norm": 0.3966093361377716,
1331
+ "learning_rate": 9.481740028736692e-05,
1332
+ "loss": 0.0318,
1333
+ "step": 1890
1334
+ },
1335
+ {
1336
+ "epoch": 0.7566706491437675,
1337
+ "grad_norm": 0.6017296314239502,
1338
+ "learning_rate": 9.474384858657164e-05,
1339
+ "loss": 0.0375,
1340
+ "step": 1900
1341
+ },
1342
+ {
1343
+ "epoch": 0.7606531262445241,
1344
+ "grad_norm": 0.6876013278961182,
1345
+ "learning_rate": 9.466980757421679e-05,
1346
+ "loss": 0.0381,
1347
+ "step": 1910
1348
+ },
1349
+ {
1350
+ "epoch": 0.7646356033452808,
1351
+ "grad_norm": 0.4606328010559082,
1352
+ "learning_rate": 9.459527806000305e-05,
1353
+ "loss": 0.0379,
1354
+ "step": 1920
1355
+ },
1356
+ {
1357
+ "epoch": 0.7686180804460374,
1358
+ "grad_norm": 0.9279587268829346,
1359
+ "learning_rate": 9.452026085897325e-05,
1360
+ "loss": 0.0351,
1361
+ "step": 1930
1362
+ },
1363
+ {
1364
+ "epoch": 0.7726005575467941,
1365
+ "grad_norm": 0.5607491135597229,
1366
+ "learning_rate": 9.444475679150348e-05,
1367
+ "loss": 0.0397,
1368
+ "step": 1940
1369
+ },
1370
+ {
1371
+ "epoch": 0.7765830346475507,
1372
+ "grad_norm": 0.7617636322975159,
1373
+ "learning_rate": 9.436876668329411e-05,
1374
+ "loss": 0.0357,
1375
+ "step": 1950
1376
+ },
1377
+ {
1378
+ "epoch": 0.7805655117483075,
1379
+ "grad_norm": 0.40550336241722107,
1380
+ "learning_rate": 9.429229136536079e-05,
1381
+ "loss": 0.0326,
1382
+ "step": 1960
1383
+ },
1384
+ {
1385
+ "epoch": 0.7845479888490641,
1386
+ "grad_norm": 0.7706924676895142,
1387
+ "learning_rate": 9.421533167402534e-05,
1388
+ "loss": 0.0395,
1389
+ "step": 1970
1390
+ },
1391
+ {
1392
+ "epoch": 0.7885304659498208,
1393
+ "grad_norm": 0.5256568789482117,
1394
+ "learning_rate": 9.413788845090666e-05,
1395
+ "loss": 0.04,
1396
+ "step": 1980
1397
+ },
1398
+ {
1399
+ "epoch": 0.7925129430505775,
1400
+ "grad_norm": 0.4805433452129364,
1401
+ "learning_rate": 9.405996254291136e-05,
1402
+ "loss": 0.0358,
1403
+ "step": 1990
1404
+ },
1405
+ {
1406
+ "epoch": 0.7964954201513341,
1407
+ "grad_norm": 0.4774327874183655,
1408
+ "learning_rate": 9.398155480222474e-05,
1409
+ "loss": 0.0323,
1410
+ "step": 2000
1411
+ },
1412
+ {
1413
+ "epoch": 0.8004778972520908,
1414
+ "grad_norm": 0.42436131834983826,
1415
+ "learning_rate": 9.390266608630128e-05,
1416
+ "loss": 0.0351,
1417
+ "step": 2010
1418
+ },
1419
+ {
1420
+ "epoch": 0.8044603743528475,
1421
+ "grad_norm": 0.43557778000831604,
1422
+ "learning_rate": 9.38232972578553e-05,
1423
+ "loss": 0.0385,
1424
+ "step": 2020
1425
+ },
1426
+ {
1427
+ "epoch": 0.8084428514536042,
1428
+ "grad_norm": 0.47707608342170715,
1429
+ "learning_rate": 9.374344918485164e-05,
1430
+ "loss": 0.0371,
1431
+ "step": 2030
1432
+ },
1433
+ {
1434
+ "epoch": 0.8124253285543608,
1435
+ "grad_norm": 0.4659009873867035,
1436
+ "learning_rate": 9.366312274049602e-05,
1437
+ "loss": 0.0407,
1438
+ "step": 2040
1439
+ },
1440
+ {
1441
+ "epoch": 0.8164078056551175,
1442
+ "grad_norm": 0.4538560211658478,
1443
+ "learning_rate": 9.358231880322554e-05,
1444
+ "loss": 0.0349,
1445
+ "step": 2050
1446
+ },
1447
+ {
1448
+ "epoch": 0.8203902827558741,
1449
+ "grad_norm": 0.6766425967216492,
1450
+ "learning_rate": 9.350103825669916e-05,
1451
+ "loss": 0.0311,
1452
+ "step": 2060
1453
+ },
1454
+ {
1455
+ "epoch": 0.8243727598566308,
1456
+ "grad_norm": 0.3502272963523865,
1457
+ "learning_rate": 9.341928198978787e-05,
1458
+ "loss": 0.0347,
1459
+ "step": 2070
1460
+ },
1461
+ {
1462
+ "epoch": 0.8283552369573874,
1463
+ "grad_norm": 0.5337536931037903,
1464
+ "learning_rate": 9.333705089656512e-05,
1465
+ "loss": 0.03,
1466
+ "step": 2080
1467
+ },
1468
+ {
1469
+ "epoch": 0.8323377140581442,
1470
+ "grad_norm": 0.6512131690979004,
1471
+ "learning_rate": 9.325434587629698e-05,
1472
+ "loss": 0.0329,
1473
+ "step": 2090
1474
+ },
1475
+ {
1476
+ "epoch": 0.8363201911589009,
1477
+ "grad_norm": 0.5638608336448669,
1478
+ "learning_rate": 9.31711678334323e-05,
1479
+ "loss": 0.0311,
1480
+ "step": 2100
1481
+ },
1482
+ {
1483
+ "epoch": 0.8403026682596575,
1484
+ "grad_norm": 0.3121340572834015,
1485
+ "learning_rate": 9.308751767759282e-05,
1486
+ "loss": 0.0319,
1487
+ "step": 2110
1488
+ },
1489
+ {
1490
+ "epoch": 0.8442851453604142,
1491
+ "grad_norm": 0.6195731163024902,
1492
+ "learning_rate": 9.300339632356325e-05,
1493
+ "loss": 0.0318,
1494
+ "step": 2120
1495
+ },
1496
+ {
1497
+ "epoch": 0.8482676224611708,
1498
+ "grad_norm": 0.6970657110214233,
1499
+ "learning_rate": 9.291880469128124e-05,
1500
+ "loss": 0.0337,
1501
+ "step": 2130
1502
+ },
1503
+ {
1504
+ "epoch": 0.8522500995619275,
1505
+ "grad_norm": 0.5600506663322449,
1506
+ "learning_rate": 9.283374370582732e-05,
1507
+ "loss": 0.0371,
1508
+ "step": 2140
1509
+ },
1510
+ {
1511
+ "epoch": 0.8562325766626842,
1512
+ "grad_norm": 0.47252482175827026,
1513
+ "learning_rate": 9.274821429741482e-05,
1514
+ "loss": 0.0399,
1515
+ "step": 2150
1516
+ },
1517
+ {
1518
+ "epoch": 0.8602150537634409,
1519
+ "grad_norm": 0.511959969997406,
1520
+ "learning_rate": 9.266221740137961e-05,
1521
+ "loss": 0.0336,
1522
+ "step": 2160
1523
+ },
1524
+ {
1525
+ "epoch": 0.8641975308641975,
1526
+ "grad_norm": 0.4836970567703247,
1527
+ "learning_rate": 9.257575395817001e-05,
1528
+ "loss": 0.0297,
1529
+ "step": 2170
1530
+ },
1531
+ {
1532
+ "epoch": 0.8681800079649542,
1533
+ "grad_norm": 0.6154841184616089,
1534
+ "learning_rate": 9.248882491333637e-05,
1535
+ "loss": 0.0369,
1536
+ "step": 2180
1537
+ },
1538
+ {
1539
+ "epoch": 0.8721624850657109,
1540
+ "grad_norm": 0.6111750602722168,
1541
+ "learning_rate": 9.240143121752076e-05,
1542
+ "loss": 0.0325,
1543
+ "step": 2190
1544
+ },
1545
+ {
1546
+ "epoch": 0.8761449621664675,
1547
+ "grad_norm": 0.4122430980205536,
1548
+ "learning_rate": 9.23135738264467e-05,
1549
+ "loss": 0.0325,
1550
+ "step": 2200
1551
+ },
1552
+ {
1553
+ "epoch": 0.8801274392672243,
1554
+ "grad_norm": 0.35100454092025757,
1555
+ "learning_rate": 9.222525370090849e-05,
1556
+ "loss": 0.0313,
1557
+ "step": 2210
1558
+ },
1559
+ {
1560
+ "epoch": 0.8841099163679809,
1561
+ "grad_norm": 0.5073012113571167,
1562
+ "learning_rate": 9.213647180676088e-05,
1563
+ "loss": 0.0337,
1564
+ "step": 2220
1565
+ },
1566
+ {
1567
+ "epoch": 0.8880923934687376,
1568
+ "grad_norm": 0.6666367650032043,
1569
+ "learning_rate": 9.204722911490846e-05,
1570
+ "loss": 0.0357,
1571
+ "step": 2230
1572
+ },
1573
+ {
1574
+ "epoch": 0.8920748705694942,
1575
+ "grad_norm": 0.43829846382141113,
1576
+ "learning_rate": 9.1957526601295e-05,
1577
+ "loss": 0.0309,
1578
+ "step": 2240
1579
+ },
1580
+ {
1581
+ "epoch": 0.8960573476702509,
1582
+ "grad_norm": 0.41505926847457886,
1583
+ "learning_rate": 9.186736524689281e-05,
1584
+ "loss": 0.0362,
1585
+ "step": 2250
1586
+ },
1587
+ {
1588
+ "epoch": 0.9000398247710075,
1589
+ "grad_norm": 0.585198700428009,
1590
+ "learning_rate": 9.177674603769204e-05,
1591
+ "loss": 0.0373,
1592
+ "step": 2260
1593
+ },
1594
+ {
1595
+ "epoch": 0.9040223018717642,
1596
+ "grad_norm": 0.5691582560539246,
1597
+ "learning_rate": 9.168566996468983e-05,
1598
+ "loss": 0.0297,
1599
+ "step": 2270
1600
+ },
1601
+ {
1602
+ "epoch": 0.9080047789725209,
1603
+ "grad_norm": 0.568215012550354,
1604
+ "learning_rate": 9.159413802387951e-05,
1605
+ "loss": 0.0375,
1606
+ "step": 2280
1607
+ },
1608
+ {
1609
+ "epoch": 0.9119872560732776,
1610
+ "grad_norm": 0.41294440627098083,
1611
+ "learning_rate": 9.150215121623974e-05,
1612
+ "loss": 0.0348,
1613
+ "step": 2290
1614
+ },
1615
+ {
1616
+ "epoch": 0.9159697331740343,
1617
+ "grad_norm": 0.5145776271820068,
1618
+ "learning_rate": 9.140971054772349e-05,
1619
+ "loss": 0.0318,
1620
+ "step": 2300
1621
+ },
1622
+ {
1623
+ "epoch": 0.9199522102747909,
1624
+ "grad_norm": 0.4824906289577484,
1625
+ "learning_rate": 9.131681702924713e-05,
1626
+ "loss": 0.0383,
1627
+ "step": 2310
1628
+ },
1629
+ {
1630
+ "epoch": 0.9239346873755476,
1631
+ "grad_norm": 0.5607543587684631,
1632
+ "learning_rate": 9.122347167667926e-05,
1633
+ "loss": 0.0288,
1634
+ "step": 2320
1635
+ },
1636
+ {
1637
+ "epoch": 0.9279171644763042,
1638
+ "grad_norm": 0.7051650285720825,
1639
+ "learning_rate": 9.112967551082973e-05,
1640
+ "loss": 0.0393,
1641
+ "step": 2330
1642
+ },
1643
+ {
1644
+ "epoch": 0.931899641577061,
1645
+ "grad_norm": 0.4049379825592041,
1646
+ "learning_rate": 9.103542955743835e-05,
1647
+ "loss": 0.0322,
1648
+ "step": 2340
1649
+ },
1650
+ {
1651
+ "epoch": 0.9358821186778176,
1652
+ "grad_norm": 0.3805215060710907,
1653
+ "learning_rate": 9.094073484716381e-05,
1654
+ "loss": 0.0357,
1655
+ "step": 2350
1656
+ },
1657
+ {
1658
+ "epoch": 0.9398645957785743,
1659
+ "grad_norm": 0.44789791107177734,
1660
+ "learning_rate": 9.084559241557226e-05,
1661
+ "loss": 0.0291,
1662
+ "step": 2360
1663
+ },
1664
+ {
1665
+ "epoch": 0.9438470728793309,
1666
+ "grad_norm": 0.6356040835380554,
1667
+ "learning_rate": 9.075000330312608e-05,
1668
+ "loss": 0.0322,
1669
+ "step": 2370
1670
+ },
1671
+ {
1672
+ "epoch": 0.9478295499800876,
1673
+ "grad_norm": 0.504687488079071,
1674
+ "learning_rate": 9.065396855517253e-05,
1675
+ "loss": 0.03,
1676
+ "step": 2380
1677
+ },
1678
+ {
1679
+ "epoch": 0.9518120270808443,
1680
+ "grad_norm": 0.4105326533317566,
1681
+ "learning_rate": 9.055748922193219e-05,
1682
+ "loss": 0.0349,
1683
+ "step": 2390
1684
+ },
1685
+ {
1686
+ "epoch": 0.955794504181601,
1687
+ "grad_norm": 0.4912046790122986,
1688
+ "learning_rate": 9.046056635848761e-05,
1689
+ "loss": 0.0384,
1690
+ "step": 2400
1691
+ },
1692
+ {
1693
+ "epoch": 0.9597769812823577,
1694
+ "grad_norm": 0.7549566030502319,
1695
+ "learning_rate": 9.036320102477169e-05,
1696
+ "loss": 0.0327,
1697
+ "step": 2410
1698
+ },
1699
+ {
1700
+ "epoch": 0.9637594583831143,
1701
+ "grad_norm": 0.36022019386291504,
1702
+ "learning_rate": 9.02653942855561e-05,
1703
+ "loss": 0.0331,
1704
+ "step": 2420
1705
+ },
1706
+ {
1707
+ "epoch": 0.967741935483871,
1708
+ "grad_norm": 0.48688042163848877,
1709
+ "learning_rate": 9.016714721043971e-05,
1710
+ "loss": 0.0318,
1711
+ "step": 2430
1712
+ },
1713
+ {
1714
+ "epoch": 0.9717244125846276,
1715
+ "grad_norm": 0.3235294222831726,
1716
+ "learning_rate": 9.006846087383675e-05,
1717
+ "loss": 0.0327,
1718
+ "step": 2440
1719
+ },
1720
+ {
1721
+ "epoch": 0.9757068896853843,
1722
+ "grad_norm": 0.5939168334007263,
1723
+ "learning_rate": 8.996933635496523e-05,
1724
+ "loss": 0.0314,
1725
+ "step": 2450
1726
+ },
1727
+ {
1728
+ "epoch": 0.9796893667861409,
1729
+ "grad_norm": 0.7387416958808899,
1730
+ "learning_rate": 8.986977473783498e-05,
1731
+ "loss": 0.0374,
1732
+ "step": 2460
1733
+ },
1734
+ {
1735
+ "epoch": 0.9836718438868977,
1736
+ "grad_norm": 0.48453521728515625,
1737
+ "learning_rate": 8.97697771112359e-05,
1738
+ "loss": 0.0317,
1739
+ "step": 2470
1740
+ },
1741
+ {
1742
+ "epoch": 0.9876543209876543,
1743
+ "grad_norm": 0.40206247568130493,
1744
+ "learning_rate": 8.966934456872602e-05,
1745
+ "loss": 0.0303,
1746
+ "step": 2480
1747
+ },
1748
+ {
1749
+ "epoch": 0.991636798088411,
1750
+ "grad_norm": 0.5897948741912842,
1751
+ "learning_rate": 8.95684782086195e-05,
1752
+ "loss": 0.0293,
1753
+ "step": 2490
1754
+ },
1755
+ {
1756
+ "epoch": 0.9956192751891677,
1757
+ "grad_norm": 0.3984488546848297,
1758
+ "learning_rate": 8.946717913397476e-05,
1759
+ "loss": 0.0374,
1760
+ "step": 2500
1761
+ },
1762
+ {
1763
+ "epoch": 0.9996017522899243,
1764
+ "grad_norm": 0.3679232597351074,
1765
+ "learning_rate": 8.93654484525822e-05,
1766
+ "loss": 0.0311,
1767
+ "step": 2510
1768
+ },
1769
+ {
1770
+ "epoch": 1.003584229390681,
1771
+ "grad_norm": 0.3514699637889862,
1772
+ "learning_rate": 8.926328727695226e-05,
1773
+ "loss": 0.0282,
1774
+ "step": 2520
1775
+ },
1776
+ {
1777
+ "epoch": 1.0075667064914378,
1778
+ "grad_norm": 0.34111514687538147,
1779
+ "learning_rate": 8.916069672430319e-05,
1780
+ "loss": 0.0272,
1781
+ "step": 2530
1782
+ },
1783
+ {
1784
+ "epoch": 1.0115491835921944,
1785
+ "grad_norm": 0.3726666569709778,
1786
+ "learning_rate": 8.905767791654884e-05,
1787
+ "loss": 0.0315,
1788
+ "step": 2540
1789
+ },
1790
+ {
1791
+ "epoch": 1.015531660692951,
1792
+ "grad_norm": 0.5761587619781494,
1793
+ "learning_rate": 8.895423198028638e-05,
1794
+ "loss": 0.0363,
1795
+ "step": 2550
1796
+ },
1797
+ {
1798
+ "epoch": 1.0195141377937076,
1799
+ "grad_norm": 0.5593805909156799,
1800
+ "learning_rate": 8.885036004678402e-05,
1801
+ "loss": 0.0311,
1802
+ "step": 2560
1803
+ },
1804
+ {
1805
+ "epoch": 1.0234966148944644,
1806
+ "grad_norm": 0.3937326669692993,
1807
+ "learning_rate": 8.874606325196857e-05,
1808
+ "loss": 0.0339,
1809
+ "step": 2570
1810
+ },
1811
+ {
1812
+ "epoch": 1.027479091995221,
1813
+ "grad_norm": 0.5169225931167603,
1814
+ "learning_rate": 8.864134273641304e-05,
1815
+ "loss": 0.03,
1816
+ "step": 2580
1817
+ },
1818
+ {
1819
+ "epoch": 1.0314615690959776,
1820
+ "grad_norm": 0.5744480490684509,
1821
+ "learning_rate": 8.853619964532427e-05,
1822
+ "loss": 0.0324,
1823
+ "step": 2590
1824
+ },
1825
+ {
1826
+ "epoch": 1.0354440461967345,
1827
+ "grad_norm": 0.555366039276123,
1828
+ "learning_rate": 8.843063512853019e-05,
1829
+ "loss": 0.0289,
1830
+ "step": 2600
1831
+ },
1832
+ {
1833
+ "epoch": 1.039426523297491,
1834
+ "grad_norm": 0.4377165734767914,
1835
+ "learning_rate": 8.832465034046749e-05,
1836
+ "loss": 0.0296,
1837
+ "step": 2610
1838
+ },
1839
+ {
1840
+ "epoch": 1.0434090003982477,
1841
+ "grad_norm": 0.5874027013778687,
1842
+ "learning_rate": 8.821824644016882e-05,
1843
+ "loss": 0.0317,
1844
+ "step": 2620
1845
+ },
1846
+ {
1847
+ "epoch": 1.0473914774990043,
1848
+ "grad_norm": 0.7405006289482117,
1849
+ "learning_rate": 8.811142459125019e-05,
1850
+ "loss": 0.0308,
1851
+ "step": 2630
1852
+ },
1853
+ {
1854
+ "epoch": 1.0513739545997611,
1855
+ "grad_norm": 0.4106619954109192,
1856
+ "learning_rate": 8.800418596189822e-05,
1857
+ "loss": 0.0252,
1858
+ "step": 2640
1859
+ },
1860
+ {
1861
+ "epoch": 1.0553564317005177,
1862
+ "grad_norm": 0.510388195514679,
1863
+ "learning_rate": 8.789653172485737e-05,
1864
+ "loss": 0.0297,
1865
+ "step": 2650
1866
+ },
1867
+ {
1868
+ "epoch": 1.0593389088012743,
1869
+ "grad_norm": 0.5874786972999573,
1870
+ "learning_rate": 8.778846305741715e-05,
1871
+ "loss": 0.0359,
1872
+ "step": 2660
1873
+ },
1874
+ {
1875
+ "epoch": 1.063321385902031,
1876
+ "grad_norm": 0.5348223447799683,
1877
+ "learning_rate": 8.767998114139918e-05,
1878
+ "loss": 0.0313,
1879
+ "step": 2670
1880
+ },
1881
+ {
1882
+ "epoch": 1.0673038630027878,
1883
+ "grad_norm": 0.3929119408130646,
1884
+ "learning_rate": 8.757108716314429e-05,
1885
+ "loss": 0.0329,
1886
+ "step": 2680
1887
+ },
1888
+ {
1889
+ "epoch": 1.0712863401035444,
1890
+ "grad_norm": 0.5612719058990479,
1891
+ "learning_rate": 8.746178231349962e-05,
1892
+ "loss": 0.0312,
1893
+ "step": 2690
1894
+ },
1895
+ {
1896
+ "epoch": 1.075268817204301,
1897
+ "grad_norm": 0.37800925970077515,
1898
+ "learning_rate": 8.735206778780549e-05,
1899
+ "loss": 0.0349,
1900
+ "step": 2700
1901
+ },
1902
+ {
1903
+ "epoch": 1.0792512943050578,
1904
+ "grad_norm": 0.43637755513191223,
1905
+ "learning_rate": 8.724194478588234e-05,
1906
+ "loss": 0.0307,
1907
+ "step": 2710
1908
+ },
1909
+ {
1910
+ "epoch": 1.0832337714058144,
1911
+ "grad_norm": 0.3608951270580292,
1912
+ "learning_rate": 8.713141451201772e-05,
1913
+ "loss": 0.0277,
1914
+ "step": 2720
1915
+ },
1916
+ {
1917
+ "epoch": 1.087216248506571,
1918
+ "grad_norm": 0.5240746140480042,
1919
+ "learning_rate": 8.702047817495295e-05,
1920
+ "loss": 0.0302,
1921
+ "step": 2730
1922
+ },
1923
+ {
1924
+ "epoch": 1.0911987256073277,
1925
+ "grad_norm": 0.3599265515804291,
1926
+ "learning_rate": 8.69091369878701e-05,
1927
+ "loss": 0.0313,
1928
+ "step": 2740
1929
+ },
1930
+ {
1931
+ "epoch": 1.0951812027080845,
1932
+ "grad_norm": 0.6075289845466614,
1933
+ "learning_rate": 8.679739216837849e-05,
1934
+ "loss": 0.0296,
1935
+ "step": 2750
1936
+ },
1937
+ {
1938
+ "epoch": 1.099163679808841,
1939
+ "grad_norm": 0.4289771318435669,
1940
+ "learning_rate": 8.66852449385016e-05,
1941
+ "loss": 0.0276,
1942
+ "step": 2760
1943
+ },
1944
+ {
1945
+ "epoch": 1.1031461569095977,
1946
+ "grad_norm": 0.3706158697605133,
1947
+ "learning_rate": 8.657269652466356e-05,
1948
+ "loss": 0.0278,
1949
+ "step": 2770
1950
+ },
1951
+ {
1952
+ "epoch": 1.1071286340103543,
1953
+ "grad_norm": 0.6201577186584473,
1954
+ "learning_rate": 8.645974815767577e-05,
1955
+ "loss": 0.0315,
1956
+ "step": 2780
1957
+ },
1958
+ {
1959
+ "epoch": 1.1111111111111112,
1960
+ "grad_norm": 0.5606251358985901,
1961
+ "learning_rate": 8.634640107272351e-05,
1962
+ "loss": 0.0274,
1963
+ "step": 2790
1964
+ },
1965
+ {
1966
+ "epoch": 1.1150935882118678,
1967
+ "grad_norm": 0.4653207063674927,
1968
+ "learning_rate": 8.623265650935234e-05,
1969
+ "loss": 0.0293,
1970
+ "step": 2800
1971
+ },
1972
+ {
1973
+ "epoch": 1.1190760653126244,
1974
+ "grad_norm": 0.4102509915828705,
1975
+ "learning_rate": 8.611851571145456e-05,
1976
+ "loss": 0.0311,
1977
+ "step": 2810
1978
+ },
1979
+ {
1980
+ "epoch": 1.1230585424133812,
1981
+ "grad_norm": 0.5217996835708618,
1982
+ "learning_rate": 8.600397992725566e-05,
1983
+ "loss": 0.0248,
1984
+ "step": 2820
1985
+ },
1986
+ {
1987
+ "epoch": 1.1270410195141378,
1988
+ "grad_norm": 0.5519301295280457,
1989
+ "learning_rate": 8.588905040930061e-05,
1990
+ "loss": 0.0314,
1991
+ "step": 2830
1992
+ },
1993
+ {
1994
+ "epoch": 1.1310234966148944,
1995
+ "grad_norm": 0.43869367241859436,
1996
+ "learning_rate": 8.577372841444022e-05,
1997
+ "loss": 0.0289,
1998
+ "step": 2840
1999
+ },
2000
+ {
2001
+ "epoch": 1.135005973715651,
2002
+ "grad_norm": 0.4673241078853607,
2003
+ "learning_rate": 8.565801520381736e-05,
2004
+ "loss": 0.026,
2005
+ "step": 2850
2006
+ },
2007
+ {
2008
+ "epoch": 1.1389884508164079,
2009
+ "grad_norm": 0.4510667622089386,
2010
+ "learning_rate": 8.554191204285313e-05,
2011
+ "loss": 0.0319,
2012
+ "step": 2860
2013
+ },
2014
+ {
2015
+ "epoch": 1.1429709279171645,
2016
+ "grad_norm": 0.473105788230896,
2017
+ "learning_rate": 8.542542020123315e-05,
2018
+ "loss": 0.032,
2019
+ "step": 2870
2020
+ },
2021
+ {
2022
+ "epoch": 1.146953405017921,
2023
+ "grad_norm": 0.4328536093235016,
2024
+ "learning_rate": 8.530854095289347e-05,
2025
+ "loss": 0.029,
2026
+ "step": 2880
2027
+ },
2028
+ {
2029
+ "epoch": 1.1509358821186777,
2030
+ "grad_norm": 0.45955684781074524,
2031
+ "learning_rate": 8.519127557600688e-05,
2032
+ "loss": 0.0274,
2033
+ "step": 2890
2034
+ },
2035
+ {
2036
+ "epoch": 1.1549183592194345,
2037
+ "grad_norm": 0.3034651577472687,
2038
+ "learning_rate": 8.507362535296871e-05,
2039
+ "loss": 0.0286,
2040
+ "step": 2900
2041
+ },
2042
+ {
2043
+ "epoch": 1.1589008363201911,
2044
+ "grad_norm": 0.6441288590431213,
2045
+ "learning_rate": 8.495559157038299e-05,
2046
+ "loss": 0.0377,
2047
+ "step": 2910
2048
+ },
2049
+ {
2050
+ "epoch": 1.1628833134209478,
2051
+ "grad_norm": 0.43017473816871643,
2052
+ "learning_rate": 8.483717551904823e-05,
2053
+ "loss": 0.0297,
2054
+ "step": 2920
2055
+ },
2056
+ {
2057
+ "epoch": 1.1668657905217046,
2058
+ "grad_norm": 0.5794181227684021,
2059
+ "learning_rate": 8.47183784939434e-05,
2060
+ "loss": 0.0281,
2061
+ "step": 2930
2062
+ },
2063
+ {
2064
+ "epoch": 1.1708482676224612,
2065
+ "grad_norm": 0.32613155245780945,
2066
+ "learning_rate": 8.459920179421374e-05,
2067
+ "loss": 0.0324,
2068
+ "step": 2940
2069
+ },
2070
+ {
2071
+ "epoch": 1.1748307447232178,
2072
+ "grad_norm": 0.514627993106842,
2073
+ "learning_rate": 8.447964672315656e-05,
2074
+ "loss": 0.0275,
2075
+ "step": 2950
2076
+ },
2077
+ {
2078
+ "epoch": 1.1788132218239746,
2079
+ "grad_norm": 0.30379730463027954,
2080
+ "learning_rate": 8.435971458820692e-05,
2081
+ "loss": 0.0322,
2082
+ "step": 2960
2083
+ },
2084
+ {
2085
+ "epoch": 1.1827956989247312,
2086
+ "grad_norm": 0.5878794193267822,
2087
+ "learning_rate": 8.423940670092345e-05,
2088
+ "loss": 0.0318,
2089
+ "step": 2970
2090
+ },
2091
+ {
2092
+ "epoch": 1.1867781760254879,
2093
+ "grad_norm": 0.3879925012588501,
2094
+ "learning_rate": 8.411872437697394e-05,
2095
+ "loss": 0.0323,
2096
+ "step": 2980
2097
+ },
2098
+ {
2099
+ "epoch": 1.1907606531262445,
2100
+ "grad_norm": 0.2640171945095062,
2101
+ "learning_rate": 8.399766893612096e-05,
2102
+ "loss": 0.0265,
2103
+ "step": 2990
2104
+ },
2105
+ {
2106
+ "epoch": 1.194743130227001,
2107
+ "grad_norm": 0.4299219846725464,
2108
+ "learning_rate": 8.38762417022074e-05,
2109
+ "loss": 0.0312,
2110
+ "step": 3000
2111
+ },
2112
+ {
2113
+ "epoch": 1.198725607327758,
2114
+ "grad_norm": 0.3073621690273285,
2115
+ "learning_rate": 8.375444400314204e-05,
2116
+ "loss": 0.0253,
2117
+ "step": 3010
2118
+ },
2119
+ {
2120
+ "epoch": 1.2027080844285145,
2121
+ "grad_norm": 0.26146814227104187,
2122
+ "learning_rate": 8.3632277170885e-05,
2123
+ "loss": 0.0312,
2124
+ "step": 3020
2125
+ },
2126
+ {
2127
+ "epoch": 1.2066905615292711,
2128
+ "grad_norm": 0.2975267767906189,
2129
+ "learning_rate": 8.350974254143318e-05,
2130
+ "loss": 0.0271,
2131
+ "step": 3030
2132
+ },
2133
+ {
2134
+ "epoch": 1.210673038630028,
2135
+ "grad_norm": 0.20422956347465515,
2136
+ "learning_rate": 8.338684145480566e-05,
2137
+ "loss": 0.0259,
2138
+ "step": 3040
2139
+ },
2140
+ {
2141
+ "epoch": 1.2146555157307846,
2142
+ "grad_norm": 0.31191202998161316,
2143
+ "learning_rate": 8.326357525502904e-05,
2144
+ "loss": 0.0304,
2145
+ "step": 3050
2146
+ },
2147
+ {
2148
+ "epoch": 1.2186379928315412,
2149
+ "grad_norm": 0.5126979947090149,
2150
+ "learning_rate": 8.313994529012273e-05,
2151
+ "loss": 0.0294,
2152
+ "step": 3060
2153
+ },
2154
+ {
2155
+ "epoch": 1.222620469932298,
2156
+ "grad_norm": 0.3219705820083618,
2157
+ "learning_rate": 8.301595291208422e-05,
2158
+ "loss": 0.0305,
2159
+ "step": 3070
2160
+ },
2161
+ {
2162
+ "epoch": 1.2266029470330546,
2163
+ "grad_norm": 0.4278619587421417,
2164
+ "learning_rate": 8.289159947687427e-05,
2165
+ "loss": 0.0297,
2166
+ "step": 3080
2167
+ },
2168
+ {
2169
+ "epoch": 1.2305854241338112,
2170
+ "grad_norm": 0.5774158239364624,
2171
+ "learning_rate": 8.276688634440216e-05,
2172
+ "loss": 0.0296,
2173
+ "step": 3090
2174
+ },
2175
+ {
2176
+ "epoch": 1.2345679012345678,
2177
+ "grad_norm": 0.5146026015281677,
2178
+ "learning_rate": 8.26418148785107e-05,
2179
+ "loss": 0.0311,
2180
+ "step": 3100
2181
+ },
2182
+ {
2183
+ "epoch": 1.2385503783353247,
2184
+ "grad_norm": 0.347504198551178,
2185
+ "learning_rate": 8.251638644696141e-05,
2186
+ "loss": 0.0345,
2187
+ "step": 3110
2188
+ },
2189
+ {
2190
+ "epoch": 1.2425328554360813,
2191
+ "grad_norm": 0.5153434872627258,
2192
+ "learning_rate": 8.23906024214195e-05,
2193
+ "loss": 0.0276,
2194
+ "step": 3120
2195
+ },
2196
+ {
2197
+ "epoch": 1.2465153325368379,
2198
+ "grad_norm": 0.31597134470939636,
2199
+ "learning_rate": 8.226446417743897e-05,
2200
+ "loss": 0.0343,
2201
+ "step": 3130
2202
+ },
2203
+ {
2204
+ "epoch": 1.2504978096375945,
2205
+ "grad_norm": 0.4005573093891144,
2206
+ "learning_rate": 8.213797309444742e-05,
2207
+ "loss": 0.0294,
2208
+ "step": 3140
2209
+ },
2210
+ {
2211
+ "epoch": 1.2544802867383513,
2212
+ "grad_norm": 0.3009648025035858,
2213
+ "learning_rate": 8.201113055573105e-05,
2214
+ "loss": 0.0287,
2215
+ "step": 3150
2216
+ },
2217
+ {
2218
+ "epoch": 1.258462763839108,
2219
+ "grad_norm": 0.6014096140861511,
2220
+ "learning_rate": 8.188393794841958e-05,
2221
+ "loss": 0.0261,
2222
+ "step": 3160
2223
+ },
2224
+ {
2225
+ "epoch": 1.2624452409398645,
2226
+ "grad_norm": 0.38871604204177856,
2227
+ "learning_rate": 8.175639666347094e-05,
2228
+ "loss": 0.0316,
2229
+ "step": 3170
2230
+ },
2231
+ {
2232
+ "epoch": 1.2664277180406214,
2233
+ "grad_norm": 0.4298296570777893,
2234
+ "learning_rate": 8.162850809565623e-05,
2235
+ "loss": 0.0248,
2236
+ "step": 3180
2237
+ },
2238
+ {
2239
+ "epoch": 1.270410195141378,
2240
+ "grad_norm": 0.3726460039615631,
2241
+ "learning_rate": 8.150027364354431e-05,
2242
+ "loss": 0.0312,
2243
+ "step": 3190
2244
+ },
2245
+ {
2246
+ "epoch": 1.2743926722421346,
2247
+ "grad_norm": 0.5679003000259399,
2248
+ "learning_rate": 8.137169470948662e-05,
2249
+ "loss": 0.0346,
2250
+ "step": 3200
2251
+ },
2252
+ {
2253
+ "epoch": 1.2783751493428912,
2254
+ "grad_norm": 0.3676661252975464,
2255
+ "learning_rate": 8.124277269960179e-05,
2256
+ "loss": 0.0298,
2257
+ "step": 3210
2258
+ },
2259
+ {
2260
+ "epoch": 1.2823576264436478,
2261
+ "grad_norm": 0.34625566005706787,
2262
+ "learning_rate": 8.111350902376023e-05,
2263
+ "loss": 0.0276,
2264
+ "step": 3220
2265
+ },
2266
+ {
2267
+ "epoch": 1.2863401035444046,
2268
+ "grad_norm": 0.37914222478866577,
2269
+ "learning_rate": 8.098390509556883e-05,
2270
+ "loss": 0.028,
2271
+ "step": 3230
2272
+ },
2273
+ {
2274
+ "epoch": 1.2903225806451613,
2275
+ "grad_norm": 0.6115908622741699,
2276
+ "learning_rate": 8.085396233235536e-05,
2277
+ "loss": 0.027,
2278
+ "step": 3240
2279
+ },
2280
+ {
2281
+ "epoch": 1.2943050577459179,
2282
+ "grad_norm": 0.3943430781364441,
2283
+ "learning_rate": 8.072368215515306e-05,
2284
+ "loss": 0.0264,
2285
+ "step": 3250
2286
+ },
2287
+ {
2288
+ "epoch": 1.2982875348466747,
2289
+ "grad_norm": 0.4377286434173584,
2290
+ "learning_rate": 8.059306598868506e-05,
2291
+ "loss": 0.0334,
2292
+ "step": 3260
2293
+ },
2294
+ {
2295
+ "epoch": 1.3022700119474313,
2296
+ "grad_norm": 0.28060171008110046,
2297
+ "learning_rate": 8.046211526134888e-05,
2298
+ "loss": 0.0284,
2299
+ "step": 3270
2300
+ },
2301
+ {
2302
+ "epoch": 1.306252489048188,
2303
+ "grad_norm": 0.4140307605266571,
2304
+ "learning_rate": 8.033083140520065e-05,
2305
+ "loss": 0.027,
2306
+ "step": 3280
2307
+ },
2308
+ {
2309
+ "epoch": 1.3102349661489447,
2310
+ "grad_norm": 0.334888756275177,
2311
+ "learning_rate": 8.019921585593962e-05,
2312
+ "loss": 0.0295,
2313
+ "step": 3290
2314
+ },
2315
+ {
2316
+ "epoch": 1.3142174432497014,
2317
+ "grad_norm": 0.5044397115707397,
2318
+ "learning_rate": 8.006727005289232e-05,
2319
+ "loss": 0.0239,
2320
+ "step": 3300
2321
+ },
2322
+ {
2323
+ "epoch": 1.318199920350458,
2324
+ "grad_norm": 0.39227598905563354,
2325
+ "learning_rate": 7.993499543899692e-05,
2326
+ "loss": 0.029,
2327
+ "step": 3310
2328
+ },
2329
+ {
2330
+ "epoch": 1.3221823974512146,
2331
+ "grad_norm": 0.5009192824363708,
2332
+ "learning_rate": 7.980239346078742e-05,
2333
+ "loss": 0.0276,
2334
+ "step": 3320
2335
+ },
2336
+ {
2337
+ "epoch": 1.3261648745519714,
2338
+ "grad_norm": 0.33442339301109314,
2339
+ "learning_rate": 7.966946556837778e-05,
2340
+ "loss": 0.025,
2341
+ "step": 3330
2342
+ },
2343
+ {
2344
+ "epoch": 1.330147351652728,
2345
+ "grad_norm": 0.44179999828338623,
2346
+ "learning_rate": 7.953621321544616e-05,
2347
+ "loss": 0.0292,
2348
+ "step": 3340
2349
+ },
2350
+ {
2351
+ "epoch": 1.3341298287534846,
2352
+ "grad_norm": 0.48963797092437744,
2353
+ "learning_rate": 7.940263785921896e-05,
2354
+ "loss": 0.0287,
2355
+ "step": 3350
2356
+ },
2357
+ {
2358
+ "epoch": 1.3381123058542412,
2359
+ "grad_norm": 0.35747164487838745,
2360
+ "learning_rate": 7.926874096045482e-05,
2361
+ "loss": 0.0257,
2362
+ "step": 3360
2363
+ },
2364
+ {
2365
+ "epoch": 1.342094782954998,
2366
+ "grad_norm": 0.6534529328346252,
2367
+ "learning_rate": 7.913452398342881e-05,
2368
+ "loss": 0.036,
2369
+ "step": 3370
2370
+ },
2371
+ {
2372
+ "epoch": 1.3460772600557547,
2373
+ "grad_norm": 0.4266543686389923,
2374
+ "learning_rate": 7.89999883959163e-05,
2375
+ "loss": 0.0307,
2376
+ "step": 3380
2377
+ },
2378
+ {
2379
+ "epoch": 1.3500597371565113,
2380
+ "grad_norm": 0.5537866950035095,
2381
+ "learning_rate": 7.886513566917687e-05,
2382
+ "loss": 0.0307,
2383
+ "step": 3390
2384
+ },
2385
+ {
2386
+ "epoch": 1.3540422142572681,
2387
+ "grad_norm": 0.35446301102638245,
2388
+ "learning_rate": 7.872996727793838e-05,
2389
+ "loss": 0.0323,
2390
+ "step": 3400
2391
+ },
2392
+ {
2393
+ "epoch": 1.3580246913580247,
2394
+ "grad_norm": 0.4526813328266144,
2395
+ "learning_rate": 7.859448470038069e-05,
2396
+ "loss": 0.0275,
2397
+ "step": 3410
2398
+ },
2399
+ {
2400
+ "epoch": 1.3620071684587813,
2401
+ "grad_norm": 0.49604639410972595,
2402
+ "learning_rate": 7.845868941811956e-05,
2403
+ "loss": 0.0317,
2404
+ "step": 3420
2405
+ },
2406
+ {
2407
+ "epoch": 1.3659896455595382,
2408
+ "grad_norm": 0.3144626319408417,
2409
+ "learning_rate": 7.832258291619043e-05,
2410
+ "loss": 0.0276,
2411
+ "step": 3430
2412
+ },
2413
+ {
2414
+ "epoch": 1.3699721226602948,
2415
+ "grad_norm": 0.5718449354171753,
2416
+ "learning_rate": 7.81861666830322e-05,
2417
+ "loss": 0.0261,
2418
+ "step": 3440
2419
+ },
2420
+ {
2421
+ "epoch": 1.3739545997610514,
2422
+ "grad_norm": 0.3567497134208679,
2423
+ "learning_rate": 7.804944221047097e-05,
2424
+ "loss": 0.027,
2425
+ "step": 3450
2426
+ },
2427
+ {
2428
+ "epoch": 1.377937076861808,
2429
+ "grad_norm": 0.5491267442703247,
2430
+ "learning_rate": 7.791241099370364e-05,
2431
+ "loss": 0.033,
2432
+ "step": 3460
2433
+ },
2434
+ {
2435
+ "epoch": 1.3819195539625646,
2436
+ "grad_norm": 0.46909114718437195,
2437
+ "learning_rate": 7.777507453128163e-05,
2438
+ "loss": 0.0298,
2439
+ "step": 3470
2440
+ },
2441
+ {
2442
+ "epoch": 1.3859020310633214,
2443
+ "grad_norm": 0.4308125376701355,
2444
+ "learning_rate": 7.763743432509451e-05,
2445
+ "loss": 0.0278,
2446
+ "step": 3480
2447
+ },
2448
+ {
2449
+ "epoch": 1.389884508164078,
2450
+ "grad_norm": 0.4042797088623047,
2451
+ "learning_rate": 7.749949188035353e-05,
2452
+ "loss": 0.0312,
2453
+ "step": 3490
2454
+ },
2455
+ {
2456
+ "epoch": 1.3938669852648347,
2457
+ "grad_norm": 0.4012243151664734,
2458
+ "learning_rate": 7.736124870557516e-05,
2459
+ "loss": 0.0296,
2460
+ "step": 3500
2461
+ },
2462
+ {
2463
+ "epoch": 1.3978494623655915,
2464
+ "grad_norm": 0.31076422333717346,
2465
+ "learning_rate": 7.722270631256459e-05,
2466
+ "loss": 0.0306,
2467
+ "step": 3510
2468
+ },
2469
+ {
2470
+ "epoch": 1.401831939466348,
2471
+ "grad_norm": 0.4105532467365265,
2472
+ "learning_rate": 7.708386621639925e-05,
2473
+ "loss": 0.0262,
2474
+ "step": 3520
2475
+ },
2476
+ {
2477
+ "epoch": 1.4058144165671047,
2478
+ "grad_norm": 0.4460810124874115,
2479
+ "learning_rate": 7.694472993541219e-05,
2480
+ "loss": 0.0269,
2481
+ "step": 3530
2482
+ },
2483
+ {
2484
+ "epoch": 1.4097968936678615,
2485
+ "grad_norm": 0.4327501952648163,
2486
+ "learning_rate": 7.680529899117547e-05,
2487
+ "loss": 0.0313,
2488
+ "step": 3540
2489
+ },
2490
+ {
2491
+ "epoch": 1.4137793707686181,
2492
+ "grad_norm": 0.4426930248737335,
2493
+ "learning_rate": 7.666557490848358e-05,
2494
+ "loss": 0.0289,
2495
+ "step": 3550
2496
+ },
2497
+ {
2498
+ "epoch": 1.4177618478693748,
2499
+ "grad_norm": 0.4433678388595581,
2500
+ "learning_rate": 7.65255592153367e-05,
2501
+ "loss": 0.0307,
2502
+ "step": 3560
2503
+ },
2504
+ {
2505
+ "epoch": 1.4217443249701314,
2506
+ "grad_norm": 0.39277946949005127,
2507
+ "learning_rate": 7.638525344292402e-05,
2508
+ "loss": 0.0321,
2509
+ "step": 3570
2510
+ },
2511
+ {
2512
+ "epoch": 1.425726802070888,
2513
+ "grad_norm": 0.4349759817123413,
2514
+ "learning_rate": 7.624465912560697e-05,
2515
+ "loss": 0.0319,
2516
+ "step": 3580
2517
+ },
2518
+ {
2519
+ "epoch": 1.4297092791716448,
2520
+ "grad_norm": 0.36243054270744324,
2521
+ "learning_rate": 7.610377780090249e-05,
2522
+ "loss": 0.025,
2523
+ "step": 3590
2524
+ },
2525
+ {
2526
+ "epoch": 1.4336917562724014,
2527
+ "grad_norm": 0.23596417903900146,
2528
+ "learning_rate": 7.596261100946618e-05,
2529
+ "loss": 0.0274,
2530
+ "step": 3600
2531
+ },
2532
+ {
2533
+ "epoch": 1.437674233373158,
2534
+ "grad_norm": 0.3828240633010864,
2535
+ "learning_rate": 7.582116029507542e-05,
2536
+ "loss": 0.0259,
2537
+ "step": 3610
2538
+ },
2539
+ {
2540
+ "epoch": 1.4416567104739149,
2541
+ "grad_norm": 0.4790317714214325,
2542
+ "learning_rate": 7.56794272046126e-05,
2543
+ "loss": 0.0278,
2544
+ "step": 3620
2545
+ },
2546
+ {
2547
+ "epoch": 1.4456391875746715,
2548
+ "grad_norm": 0.46134695410728455,
2549
+ "learning_rate": 7.55374132880481e-05,
2550
+ "loss": 0.0298,
2551
+ "step": 3630
2552
+ },
2553
+ {
2554
+ "epoch": 1.449621664675428,
2555
+ "grad_norm": 0.4513108730316162,
2556
+ "learning_rate": 7.539512009842333e-05,
2557
+ "loss": 0.0313,
2558
+ "step": 3640
2559
+ },
2560
+ {
2561
+ "epoch": 1.453604141776185,
2562
+ "grad_norm": 0.5272055268287659,
2563
+ "learning_rate": 7.525254919183382e-05,
2564
+ "loss": 0.0283,
2565
+ "step": 3650
2566
+ },
2567
+ {
2568
+ "epoch": 1.4575866188769415,
2569
+ "grad_norm": 0.43485143780708313,
2570
+ "learning_rate": 7.510970212741215e-05,
2571
+ "loss": 0.0272,
2572
+ "step": 3660
2573
+ },
2574
+ {
2575
+ "epoch": 1.4615690959776981,
2576
+ "grad_norm": 0.44522616267204285,
2577
+ "learning_rate": 7.496658046731096e-05,
2578
+ "loss": 0.0281,
2579
+ "step": 3670
2580
+ },
2581
+ {
2582
+ "epoch": 1.4655515730784547,
2583
+ "grad_norm": 0.7030194997787476,
2584
+ "learning_rate": 7.482318577668578e-05,
2585
+ "loss": 0.0336,
2586
+ "step": 3680
2587
+ },
2588
+ {
2589
+ "epoch": 1.4695340501792113,
2590
+ "grad_norm": 0.47834673523902893,
2591
+ "learning_rate": 7.467951962367796e-05,
2592
+ "loss": 0.0276,
2593
+ "step": 3690
2594
+ },
2595
+ {
2596
+ "epoch": 1.4735165272799682,
2597
+ "grad_norm": 0.4326513409614563,
2598
+ "learning_rate": 7.453558357939755e-05,
2599
+ "loss": 0.0268,
2600
+ "step": 3700
2601
+ },
2602
+ {
2603
+ "epoch": 1.4774990043807248,
2604
+ "grad_norm": 0.4779656231403351,
2605
+ "learning_rate": 7.439137921790606e-05,
2606
+ "loss": 0.0322,
2607
+ "step": 3710
2608
+ },
2609
+ {
2610
+ "epoch": 1.4814814814814814,
2611
+ "grad_norm": 0.5998985767364502,
2612
+ "learning_rate": 7.42469081161993e-05,
2613
+ "loss": 0.0257,
2614
+ "step": 3720
2615
+ },
2616
+ {
2617
+ "epoch": 1.4854639585822382,
2618
+ "grad_norm": 0.44799843430519104,
2619
+ "learning_rate": 7.410217185419006e-05,
2620
+ "loss": 0.0235,
2621
+ "step": 3730
2622
+ },
2623
+ {
2624
+ "epoch": 1.4894464356829948,
2625
+ "grad_norm": 0.335964560508728,
2626
+ "learning_rate": 7.395717201469095e-05,
2627
+ "loss": 0.0286,
2628
+ "step": 3740
2629
+ },
2630
+ {
2631
+ "epoch": 1.4934289127837514,
2632
+ "grad_norm": 0.5583806037902832,
2633
+ "learning_rate": 7.381191018339696e-05,
2634
+ "loss": 0.0292,
2635
+ "step": 3750
2636
+ },
2637
+ {
2638
+ "epoch": 1.4974113898845083,
2639
+ "grad_norm": 0.40853527188301086,
2640
+ "learning_rate": 7.36663879488682e-05,
2641
+ "loss": 0.024,
2642
+ "step": 3760
2643
+ },
2644
+ {
2645
+ "epoch": 1.5013938669852647,
2646
+ "grad_norm": 0.5366830825805664,
2647
+ "learning_rate": 7.352060690251254e-05,
2648
+ "loss": 0.0278,
2649
+ "step": 3770
2650
+ },
2651
+ {
2652
+ "epoch": 1.5053763440860215,
2653
+ "grad_norm": 0.5075726509094238,
2654
+ "learning_rate": 7.337456863856811e-05,
2655
+ "loss": 0.031,
2656
+ "step": 3780
2657
+ },
2658
+ {
2659
+ "epoch": 1.5093588211867783,
2660
+ "grad_norm": 0.4533558487892151,
2661
+ "learning_rate": 7.3228274754086e-05,
2662
+ "loss": 0.0232,
2663
+ "step": 3790
2664
+ },
2665
+ {
2666
+ "epoch": 1.5133412982875347,
2667
+ "grad_norm": 0.4715867340564728,
2668
+ "learning_rate": 7.308172684891267e-05,
2669
+ "loss": 0.0319,
2670
+ "step": 3800
2671
+ },
2672
+ {
2673
+ "epoch": 1.5173237753882916,
2674
+ "grad_norm": 0.2563737630844116,
2675
+ "learning_rate": 7.293492652567255e-05,
2676
+ "loss": 0.0239,
2677
+ "step": 3810
2678
+ },
2679
+ {
2680
+ "epoch": 1.5213062524890482,
2681
+ "grad_norm": 0.3013695478439331,
2682
+ "learning_rate": 7.278787538975043e-05,
2683
+ "loss": 0.0302,
2684
+ "step": 3820
2685
+ },
2686
+ {
2687
+ "epoch": 1.5252887295898048,
2688
+ "grad_norm": 0.39960938692092896,
2689
+ "learning_rate": 7.2640575049274e-05,
2690
+ "loss": 0.0251,
2691
+ "step": 3830
2692
+ },
2693
+ {
2694
+ "epoch": 1.5292712066905616,
2695
+ "grad_norm": 0.380015105009079,
2696
+ "learning_rate": 7.249302711509616e-05,
2697
+ "loss": 0.0231,
2698
+ "step": 3840
2699
+ },
2700
+ {
2701
+ "epoch": 1.5332536837913182,
2702
+ "grad_norm": 0.2648374140262604,
2703
+ "learning_rate": 7.23452332007775e-05,
2704
+ "loss": 0.0225,
2705
+ "step": 3850
2706
+ },
2707
+ {
2708
+ "epoch": 1.5372361608920748,
2709
+ "grad_norm": 0.32741981744766235,
2710
+ "learning_rate": 7.219719492256858e-05,
2711
+ "loss": 0.0272,
2712
+ "step": 3860
2713
+ },
2714
+ {
2715
+ "epoch": 1.5412186379928317,
2716
+ "grad_norm": 0.5281572341918945,
2717
+ "learning_rate": 7.20489138993923e-05,
2718
+ "loss": 0.025,
2719
+ "step": 3870
2720
+ },
2721
+ {
2722
+ "epoch": 1.5452011150935883,
2723
+ "grad_norm": 0.3067561388015747,
2724
+ "learning_rate": 7.190039175282614e-05,
2725
+ "loss": 0.0261,
2726
+ "step": 3880
2727
+ },
2728
+ {
2729
+ "epoch": 1.5491835921943449,
2730
+ "grad_norm": 0.4421009421348572,
2731
+ "learning_rate": 7.175163010708455e-05,
2732
+ "loss": 0.0245,
2733
+ "step": 3890
2734
+ },
2735
+ {
2736
+ "epoch": 1.5531660692951017,
2737
+ "grad_norm": 0.3433239758014679,
2738
+ "learning_rate": 7.1602630589001e-05,
2739
+ "loss": 0.027,
2740
+ "step": 3900
2741
+ },
2742
+ {
2743
+ "epoch": 1.557148546395858,
2744
+ "grad_norm": 0.30050766468048096,
2745
+ "learning_rate": 7.14533948280104e-05,
2746
+ "loss": 0.0269,
2747
+ "step": 3910
2748
+ },
2749
+ {
2750
+ "epoch": 1.561131023496615,
2751
+ "grad_norm": 0.3128851056098938,
2752
+ "learning_rate": 7.130392445613109e-05,
2753
+ "loss": 0.03,
2754
+ "step": 3920
2755
+ },
2756
+ {
2757
+ "epoch": 1.5651135005973715,
2758
+ "grad_norm": 0.3643544912338257,
2759
+ "learning_rate": 7.115422110794711e-05,
2760
+ "loss": 0.0278,
2761
+ "step": 3930
2762
+ },
2763
+ {
2764
+ "epoch": 1.5690959776981281,
2765
+ "grad_norm": 0.4426794648170471,
2766
+ "learning_rate": 7.100428642059033e-05,
2767
+ "loss": 0.0303,
2768
+ "step": 3940
2769
+ },
2770
+ {
2771
+ "epoch": 1.573078454798885,
2772
+ "grad_norm": 0.3874496817588806,
2773
+ "learning_rate": 7.08541220337224e-05,
2774
+ "loss": 0.0277,
2775
+ "step": 3950
2776
+ },
2777
+ {
2778
+ "epoch": 1.5770609318996416,
2779
+ "grad_norm": 0.3397146761417389,
2780
+ "learning_rate": 7.070372958951706e-05,
2781
+ "loss": 0.0271,
2782
+ "step": 3960
2783
+ },
2784
+ {
2785
+ "epoch": 1.5810434090003982,
2786
+ "grad_norm": 0.6062551736831665,
2787
+ "learning_rate": 7.055311073264194e-05,
2788
+ "loss": 0.0334,
2789
+ "step": 3970
2790
+ },
2791
+ {
2792
+ "epoch": 1.585025886101155,
2793
+ "grad_norm": 0.2893720269203186,
2794
+ "learning_rate": 7.040226711024077e-05,
2795
+ "loss": 0.0276,
2796
+ "step": 3980
2797
+ },
2798
+ {
2799
+ "epoch": 1.5890083632019116,
2800
+ "grad_norm": 0.28936243057250977,
2801
+ "learning_rate": 7.02512003719152e-05,
2802
+ "loss": 0.0239,
2803
+ "step": 3990
2804
+ },
2805
+ {
2806
+ "epoch": 1.5929908403026682,
2807
+ "grad_norm": 0.4626624286174774,
2808
+ "learning_rate": 7.00999121697069e-05,
2809
+ "loss": 0.0304,
2810
+ "step": 4000
2811
+ },
2812
+ {
2813
+ "epoch": 1.596973317403425,
2814
+ "grad_norm": 0.4734528362751007,
2815
+ "learning_rate": 6.99484041580794e-05,
2816
+ "loss": 0.0228,
2817
+ "step": 4010
2818
+ },
2819
+ {
2820
+ "epoch": 1.6009557945041815,
2821
+ "grad_norm": 0.6309609413146973,
2822
+ "learning_rate": 6.979667799390004e-05,
2823
+ "loss": 0.026,
2824
+ "step": 4020
2825
+ },
2826
+ {
2827
+ "epoch": 1.6049382716049383,
2828
+ "grad_norm": 0.4279467463493347,
2829
+ "learning_rate": 6.964473533642185e-05,
2830
+ "loss": 0.0298,
2831
+ "step": 4030
2832
+ },
2833
+ {
2834
+ "epoch": 1.608920748705695,
2835
+ "grad_norm": 0.29877039790153503,
2836
+ "learning_rate": 6.949257784726539e-05,
2837
+ "loss": 0.0328,
2838
+ "step": 4040
2839
+ },
2840
+ {
2841
+ "epoch": 1.6129032258064515,
2842
+ "grad_norm": 0.48857632279396057,
2843
+ "learning_rate": 6.934020719040056e-05,
2844
+ "loss": 0.0262,
2845
+ "step": 4050
2846
+ },
2847
+ {
2848
+ "epoch": 1.6168857029072083,
2849
+ "grad_norm": 0.5152528882026672,
2850
+ "learning_rate": 6.918762503212848e-05,
2851
+ "loss": 0.0283,
2852
+ "step": 4060
2853
+ },
2854
+ {
2855
+ "epoch": 1.620868180007965,
2856
+ "grad_norm": 0.3721228241920471,
2857
+ "learning_rate": 6.903483304106319e-05,
2858
+ "loss": 0.0265,
2859
+ "step": 4070
2860
+ },
2861
+ {
2862
+ "epoch": 1.6248506571087216,
2863
+ "grad_norm": 0.3533932864665985,
2864
+ "learning_rate": 6.888183288811341e-05,
2865
+ "loss": 0.0258,
2866
+ "step": 4080
2867
+ },
2868
+ {
2869
+ "epoch": 1.6288331342094784,
2870
+ "grad_norm": 0.2944384217262268,
2871
+ "learning_rate": 6.87286262464643e-05,
2872
+ "loss": 0.0244,
2873
+ "step": 4090
2874
+ },
2875
+ {
2876
+ "epoch": 1.632815611310235,
2877
+ "grad_norm": 0.2960841953754425,
2878
+ "learning_rate": 6.857521479155915e-05,
2879
+ "loss": 0.0287,
2880
+ "step": 4100
2881
+ },
2882
+ {
2883
+ "epoch": 1.6367980884109916,
2884
+ "grad_norm": 0.494831383228302,
2885
+ "learning_rate": 6.842160020108104e-05,
2886
+ "loss": 0.026,
2887
+ "step": 4110
2888
+ },
2889
+ {
2890
+ "epoch": 1.6407805655117484,
2891
+ "grad_norm": 0.652524471282959,
2892
+ "learning_rate": 6.826778415493455e-05,
2893
+ "loss": 0.0254,
2894
+ "step": 4120
2895
+ },
2896
+ {
2897
+ "epoch": 1.6447630426125048,
2898
+ "grad_norm": 0.315049409866333,
2899
+ "learning_rate": 6.811376833522729e-05,
2900
+ "loss": 0.0251,
2901
+ "step": 4130
2902
+ },
2903
+ {
2904
+ "epoch": 1.6487455197132617,
2905
+ "grad_norm": 0.3198871910572052,
2906
+ "learning_rate": 6.795955442625159e-05,
2907
+ "loss": 0.0277,
2908
+ "step": 4140
2909
+ },
2910
+ {
2911
+ "epoch": 1.6527279968140183,
2912
+ "grad_norm": 0.523373544216156,
2913
+ "learning_rate": 6.780514411446608e-05,
2914
+ "loss": 0.0255,
2915
+ "step": 4150
2916
+ },
2917
+ {
2918
+ "epoch": 1.6567104739147749,
2919
+ "grad_norm": 0.4847145080566406,
2920
+ "learning_rate": 6.765053908847716e-05,
2921
+ "loss": 0.0197,
2922
+ "step": 4160
2923
+ },
2924
+ {
2925
+ "epoch": 1.6606929510155317,
2926
+ "grad_norm": 0.4315938353538513,
2927
+ "learning_rate": 6.749574103902064e-05,
2928
+ "loss": 0.027,
2929
+ "step": 4170
2930
+ },
2931
+ {
2932
+ "epoch": 1.6646754281162883,
2933
+ "grad_norm": 0.3840107023715973,
2934
+ "learning_rate": 6.734075165894317e-05,
2935
+ "loss": 0.0286,
2936
+ "step": 4180
2937
+ },
2938
+ {
2939
+ "epoch": 1.668657905217045,
2940
+ "grad_norm": 0.42311540246009827,
2941
+ "learning_rate": 6.71855726431838e-05,
2942
+ "loss": 0.028,
2943
+ "step": 4190
2944
+ },
2945
+ {
2946
+ "epoch": 1.6726403823178018,
2947
+ "grad_norm": 0.38411083817481995,
2948
+ "learning_rate": 6.703020568875538e-05,
2949
+ "loss": 0.0263,
2950
+ "step": 4200
2951
+ },
2952
+ {
2953
+ "epoch": 1.6766228594185584,
2954
+ "grad_norm": 0.368691623210907,
2955
+ "learning_rate": 6.687465249472603e-05,
2956
+ "loss": 0.0266,
2957
+ "step": 4210
2958
+ },
2959
+ {
2960
+ "epoch": 1.680605336519315,
2961
+ "grad_norm": 0.5589401125907898,
2962
+ "learning_rate": 6.671891476220055e-05,
2963
+ "loss": 0.0255,
2964
+ "step": 4220
2965
+ },
2966
+ {
2967
+ "epoch": 1.6845878136200718,
2968
+ "grad_norm": 0.5339823365211487,
2969
+ "learning_rate": 6.656299419430183e-05,
2970
+ "loss": 0.0248,
2971
+ "step": 4230
2972
+ },
2973
+ {
2974
+ "epoch": 1.6885702907208282,
2975
+ "grad_norm": 0.3141709566116333,
2976
+ "learning_rate": 6.640689249615223e-05,
2977
+ "loss": 0.0261,
2978
+ "step": 4240
2979
+ },
2980
+ {
2981
+ "epoch": 1.692552767821585,
2982
+ "grad_norm": 0.31322765350341797,
2983
+ "learning_rate": 6.625061137485491e-05,
2984
+ "loss": 0.0275,
2985
+ "step": 4250
2986
+ },
2987
+ {
2988
+ "epoch": 1.6965352449223416,
2989
+ "grad_norm": 0.35308682918548584,
2990
+ "learning_rate": 6.609415253947517e-05,
2991
+ "loss": 0.0288,
2992
+ "step": 4260
2993
+ },
2994
+ {
2995
+ "epoch": 1.7005177220230983,
2996
+ "grad_norm": 0.32278379797935486,
2997
+ "learning_rate": 6.593751770102178e-05,
2998
+ "loss": 0.0214,
2999
+ "step": 4270
3000
+ },
3001
+ {
3002
+ "epoch": 1.704500199123855,
3003
+ "grad_norm": 0.2799319922924042,
3004
+ "learning_rate": 6.578070857242823e-05,
3005
+ "loss": 0.0246,
3006
+ "step": 4280
3007
+ },
3008
+ {
3009
+ "epoch": 1.7084826762246117,
3010
+ "grad_norm": 0.4078690707683563,
3011
+ "learning_rate": 6.562372686853402e-05,
3012
+ "loss": 0.022,
3013
+ "step": 4290
3014
+ },
3015
+ {
3016
+ "epoch": 1.7124651533253683,
3017
+ "grad_norm": 0.32824915647506714,
3018
+ "learning_rate": 6.546657430606593e-05,
3019
+ "loss": 0.0226,
3020
+ "step": 4300
3021
+ },
3022
+ {
3023
+ "epoch": 1.7164476304261251,
3024
+ "grad_norm": 0.4191490411758423,
3025
+ "learning_rate": 6.530925260361918e-05,
3026
+ "loss": 0.0237,
3027
+ "step": 4310
3028
+ },
3029
+ {
3030
+ "epoch": 1.7204301075268817,
3031
+ "grad_norm": 0.37340426445007324,
3032
+ "learning_rate": 6.515176348163871e-05,
3033
+ "loss": 0.026,
3034
+ "step": 4320
3035
+ },
3036
+ {
3037
+ "epoch": 1.7244125846276384,
3038
+ "grad_norm": 0.44565045833587646,
3039
+ "learning_rate": 6.499410866240032e-05,
3040
+ "loss": 0.0217,
3041
+ "step": 4330
3042
+ },
3043
+ {
3044
+ "epoch": 1.7283950617283952,
3045
+ "grad_norm": 0.37896549701690674,
3046
+ "learning_rate": 6.48362898699919e-05,
3047
+ "loss": 0.0215,
3048
+ "step": 4340
3049
+ },
3050
+ {
3051
+ "epoch": 1.7323775388291516,
3052
+ "grad_norm": 0.42717358469963074,
3053
+ "learning_rate": 6.467830883029443e-05,
3054
+ "loss": 0.0241,
3055
+ "step": 4350
3056
+ },
3057
+ {
3058
+ "epoch": 1.7363600159299084,
3059
+ "grad_norm": 0.5169128179550171,
3060
+ "learning_rate": 6.452016727096326e-05,
3061
+ "loss": 0.0244,
3062
+ "step": 4360
3063
+ },
3064
+ {
3065
+ "epoch": 1.7403424930306652,
3066
+ "grad_norm": 0.4041356146335602,
3067
+ "learning_rate": 6.436186692140916e-05,
3068
+ "loss": 0.0234,
3069
+ "step": 4370
3070
+ },
3071
+ {
3072
+ "epoch": 1.7443249701314216,
3073
+ "grad_norm": 0.41365042328834534,
3074
+ "learning_rate": 6.420340951277938e-05,
3075
+ "loss": 0.0214,
3076
+ "step": 4380
3077
+ },
3078
+ {
3079
+ "epoch": 1.7483074472321785,
3080
+ "grad_norm": 0.5286877155303955,
3081
+ "learning_rate": 6.404479677793874e-05,
3082
+ "loss": 0.0259,
3083
+ "step": 4390
3084
+ },
3085
+ {
3086
+ "epoch": 1.752289924332935,
3087
+ "grad_norm": 0.3863738775253296,
3088
+ "learning_rate": 6.388603045145075e-05,
3089
+ "loss": 0.0271,
3090
+ "step": 4400
3091
+ },
3092
+ {
3093
+ "epoch": 1.7562724014336917,
3094
+ "grad_norm": 0.3407500088214874,
3095
+ "learning_rate": 6.372711226955843e-05,
3096
+ "loss": 0.0248,
3097
+ "step": 4410
3098
+ },
3099
+ {
3100
+ "epoch": 1.7602548785344485,
3101
+ "grad_norm": 0.4132188558578491,
3102
+ "learning_rate": 6.356804397016564e-05,
3103
+ "loss": 0.0232,
3104
+ "step": 4420
3105
+ },
3106
+ {
3107
+ "epoch": 1.7642373556352051,
3108
+ "grad_norm": 0.5538100004196167,
3109
+ "learning_rate": 6.340882729281779e-05,
3110
+ "loss": 0.0257,
3111
+ "step": 4430
3112
+ },
3113
+ {
3114
+ "epoch": 1.7682198327359617,
3115
+ "grad_norm": 0.5813969969749451,
3116
+ "learning_rate": 6.324946397868294e-05,
3117
+ "loss": 0.0289,
3118
+ "step": 4440
3119
+ },
3120
+ {
3121
+ "epoch": 1.7722023098367186,
3122
+ "grad_norm": 0.381994366645813,
3123
+ "learning_rate": 6.308995577053276e-05,
3124
+ "loss": 0.0241,
3125
+ "step": 4450
3126
+ },
3127
+ {
3128
+ "epoch": 1.776184786937475,
3129
+ "grad_norm": 0.49612873792648315,
3130
+ "learning_rate": 6.293030441272347e-05,
3131
+ "loss": 0.0267,
3132
+ "step": 4460
3133
+ },
3134
+ {
3135
+ "epoch": 1.7801672640382318,
3136
+ "grad_norm": 0.5249873995780945,
3137
+ "learning_rate": 6.277051165117677e-05,
3138
+ "loss": 0.0246,
3139
+ "step": 4470
3140
+ },
3141
+ {
3142
+ "epoch": 1.7841497411389886,
3143
+ "grad_norm": 0.3832807242870331,
3144
+ "learning_rate": 6.261057923336064e-05,
3145
+ "loss": 0.0267,
3146
+ "step": 4480
3147
+ },
3148
+ {
3149
+ "epoch": 1.788132218239745,
3150
+ "grad_norm": 0.31773999333381653,
3151
+ "learning_rate": 6.245050890827042e-05,
3152
+ "loss": 0.0276,
3153
+ "step": 4490
3154
+ },
3155
+ {
3156
+ "epoch": 1.7921146953405018,
3157
+ "grad_norm": 0.275312215089798,
3158
+ "learning_rate": 6.229030242640952e-05,
3159
+ "loss": 0.0238,
3160
+ "step": 4500
3161
+ },
3162
+ {
3163
+ "epoch": 1.7960971724412584,
3164
+ "grad_norm": 0.5737529397010803,
3165
+ "learning_rate": 6.212996153977037e-05,
3166
+ "loss": 0.0308,
3167
+ "step": 4510
3168
+ },
3169
+ {
3170
+ "epoch": 1.800079649542015,
3171
+ "grad_norm": 0.33575618267059326,
3172
+ "learning_rate": 6.196948800181523e-05,
3173
+ "loss": 0.0258,
3174
+ "step": 4520
3175
+ },
3176
+ {
3177
+ "epoch": 1.8040621266427719,
3178
+ "grad_norm": 0.6271135807037354,
3179
+ "learning_rate": 6.180888356745695e-05,
3180
+ "loss": 0.0267,
3181
+ "step": 4530
3182
+ },
3183
+ {
3184
+ "epoch": 1.8080446037435285,
3185
+ "grad_norm": 0.5151987075805664,
3186
+ "learning_rate": 6.164814999303995e-05,
3187
+ "loss": 0.0239,
3188
+ "step": 4540
3189
+ },
3190
+ {
3191
+ "epoch": 1.812027080844285,
3192
+ "grad_norm": 0.26632851362228394,
3193
+ "learning_rate": 6.148728903632081e-05,
3194
+ "loss": 0.0222,
3195
+ "step": 4550
3196
+ },
3197
+ {
3198
+ "epoch": 1.816009557945042,
3199
+ "grad_norm": 0.4331364929676056,
3200
+ "learning_rate": 6.132630245644921e-05,
3201
+ "loss": 0.0226,
3202
+ "step": 4560
3203
+ },
3204
+ {
3205
+ "epoch": 1.8199920350457985,
3206
+ "grad_norm": 0.26644977927207947,
3207
+ "learning_rate": 6.116519201394857e-05,
3208
+ "loss": 0.0242,
3209
+ "step": 4570
3210
+ },
3211
+ {
3212
+ "epoch": 1.8239745121465551,
3213
+ "grad_norm": 0.45340025424957275,
3214
+ "learning_rate": 6.10039594706969e-05,
3215
+ "loss": 0.0233,
3216
+ "step": 4580
3217
+ },
3218
+ {
3219
+ "epoch": 1.827956989247312,
3220
+ "grad_norm": 0.3398999869823456,
3221
+ "learning_rate": 6.084260658990744e-05,
3222
+ "loss": 0.0226,
3223
+ "step": 4590
3224
+ },
3225
+ {
3226
+ "epoch": 1.8319394663480684,
3227
+ "grad_norm": 0.6932306885719299,
3228
+ "learning_rate": 6.068113513610943e-05,
3229
+ "loss": 0.0239,
3230
+ "step": 4600
3231
+ },
3232
+ {
3233
+ "epoch": 1.8359219434488252,
3234
+ "grad_norm": 0.23555903136730194,
3235
+ "learning_rate": 6.0519546875128876e-05,
3236
+ "loss": 0.0249,
3237
+ "step": 4610
3238
+ },
3239
+ {
3240
+ "epoch": 1.8399044205495818,
3241
+ "grad_norm": 0.49692243337631226,
3242
+ "learning_rate": 6.035784357406906e-05,
3243
+ "loss": 0.0241,
3244
+ "step": 4620
3245
+ },
3246
+ {
3247
+ "epoch": 1.8438868976503384,
3248
+ "grad_norm": 0.32976606488227844,
3249
+ "learning_rate": 6.01960270012914e-05,
3250
+ "loss": 0.0237,
3251
+ "step": 4630
3252
+ },
3253
+ {
3254
+ "epoch": 1.8478693747510953,
3255
+ "grad_norm": 0.4834267199039459,
3256
+ "learning_rate": 6.003409892639599e-05,
3257
+ "loss": 0.027,
3258
+ "step": 4640
3259
+ },
3260
+ {
3261
+ "epoch": 1.8518518518518519,
3262
+ "grad_norm": 0.4596833884716034,
3263
+ "learning_rate": 5.9872061120202336e-05,
3264
+ "loss": 0.0246,
3265
+ "step": 4650
3266
+ },
3267
+ {
3268
+ "epoch": 1.8558343289526085,
3269
+ "grad_norm": 0.427120178937912,
3270
+ "learning_rate": 5.9709915354729914e-05,
3271
+ "loss": 0.0269,
3272
+ "step": 4660
3273
+ },
3274
+ {
3275
+ "epoch": 1.8598168060533653,
3276
+ "grad_norm": 0.2686680257320404,
3277
+ "learning_rate": 5.9547663403178824e-05,
3278
+ "loss": 0.0239,
3279
+ "step": 4670
3280
+ },
3281
+ {
3282
+ "epoch": 1.863799283154122,
3283
+ "grad_norm": 0.3082379698753357,
3284
+ "learning_rate": 5.9385307039910445e-05,
3285
+ "loss": 0.0204,
3286
+ "step": 4680
3287
+ },
3288
+ {
3289
+ "epoch": 1.8677817602548785,
3290
+ "grad_norm": 0.34263017773628235,
3291
+ "learning_rate": 5.922284804042792e-05,
3292
+ "loss": 0.0249,
3293
+ "step": 4690
3294
+ },
3295
+ {
3296
+ "epoch": 1.8717642373556354,
3297
+ "grad_norm": 0.23308919370174408,
3298
+ "learning_rate": 5.906028818135687e-05,
3299
+ "loss": 0.0248,
3300
+ "step": 4700
3301
+ },
3302
+ {
3303
+ "epoch": 1.8757467144563917,
3304
+ "grad_norm": 0.29352062940597534,
3305
+ "learning_rate": 5.889762924042585e-05,
3306
+ "loss": 0.0231,
3307
+ "step": 4710
3308
+ },
3309
+ {
3310
+ "epoch": 1.8797291915571486,
3311
+ "grad_norm": 0.45019692182540894,
3312
+ "learning_rate": 5.873487299644699e-05,
3313
+ "loss": 0.0265,
3314
+ "step": 4720
3315
+ },
3316
+ {
3317
+ "epoch": 1.8837116686579052,
3318
+ "grad_norm": 0.4193218946456909,
3319
+ "learning_rate": 5.857202122929649e-05,
3320
+ "loss": 0.0228,
3321
+ "step": 4730
3322
+ },
3323
+ {
3324
+ "epoch": 1.8876941457586618,
3325
+ "grad_norm": 0.5292001366615295,
3326
+ "learning_rate": 5.840907571989518e-05,
3327
+ "loss": 0.0218,
3328
+ "step": 4740
3329
+ },
3330
+ {
3331
+ "epoch": 1.8916766228594186,
3332
+ "grad_norm": 0.36202308535575867,
3333
+ "learning_rate": 5.824603825018904e-05,
3334
+ "loss": 0.0229,
3335
+ "step": 4750
3336
+ },
3337
+ {
3338
+ "epoch": 1.8956590999601752,
3339
+ "grad_norm": 0.2474525272846222,
3340
+ "learning_rate": 5.808291060312975e-05,
3341
+ "loss": 0.0259,
3342
+ "step": 4760
3343
+ },
3344
+ {
3345
+ "epoch": 1.8996415770609318,
3346
+ "grad_norm": 0.23628737032413483,
3347
+ "learning_rate": 5.7919694562655083e-05,
3348
+ "loss": 0.0218,
3349
+ "step": 4770
3350
+ },
3351
+ {
3352
+ "epoch": 1.9036240541616887,
3353
+ "grad_norm": 0.38689592480659485,
3354
+ "learning_rate": 5.775639191366954e-05,
3355
+ "loss": 0.0246,
3356
+ "step": 4780
3357
+ },
3358
+ {
3359
+ "epoch": 1.9076065312624453,
3360
+ "grad_norm": 0.3524068593978882,
3361
+ "learning_rate": 5.75930044420247e-05,
3362
+ "loss": 0.0209,
3363
+ "step": 4790
3364
+ },
3365
+ {
3366
+ "epoch": 1.911589008363202,
3367
+ "grad_norm": 0.3510373830795288,
3368
+ "learning_rate": 5.74295339344998e-05,
3369
+ "loss": 0.026,
3370
+ "step": 4800
3371
+ },
3372
+ {
3373
+ "epoch": 1.9155714854639587,
3374
+ "grad_norm": 0.5862729549407959,
3375
+ "learning_rate": 5.726598217878211e-05,
3376
+ "loss": 0.0229,
3377
+ "step": 4810
3378
+ },
3379
+ {
3380
+ "epoch": 1.919553962564715,
3381
+ "grad_norm": 0.47685012221336365,
3382
+ "learning_rate": 5.71023509634474e-05,
3383
+ "loss": 0.0229,
3384
+ "step": 4820
3385
+ },
3386
+ {
3387
+ "epoch": 1.923536439665472,
3388
+ "grad_norm": 0.5108020901679993,
3389
+ "learning_rate": 5.693864207794049e-05,
3390
+ "loss": 0.0281,
3391
+ "step": 4830
3392
+ },
3393
+ {
3394
+ "epoch": 1.9275189167662286,
3395
+ "grad_norm": 0.45045313239097595,
3396
+ "learning_rate": 5.677485731255545e-05,
3397
+ "loss": 0.0287,
3398
+ "step": 4840
3399
+ },
3400
+ {
3401
+ "epoch": 1.9315013938669852,
3402
+ "grad_norm": 0.3624630868434906,
3403
+ "learning_rate": 5.6610998458416296e-05,
3404
+ "loss": 0.027,
3405
+ "step": 4850
3406
+ },
3407
+ {
3408
+ "epoch": 1.935483870967742,
3409
+ "grad_norm": 0.3359379768371582,
3410
+ "learning_rate": 5.644706730745716e-05,
3411
+ "loss": 0.0205,
3412
+ "step": 4860
3413
+ },
3414
+ {
3415
+ "epoch": 1.9394663480684986,
3416
+ "grad_norm": 0.2658320963382721,
3417
+ "learning_rate": 5.628306565240287e-05,
3418
+ "loss": 0.0222,
3419
+ "step": 4870
3420
+ },
3421
+ {
3422
+ "epoch": 1.9434488251692552,
3423
+ "grad_norm": 0.16451801359653473,
3424
+ "learning_rate": 5.611899528674923e-05,
3425
+ "loss": 0.0249,
3426
+ "step": 4880
3427
+ },
3428
+ {
3429
+ "epoch": 1.947431302270012,
3430
+ "grad_norm": 0.24440331757068634,
3431
+ "learning_rate": 5.595485800474349e-05,
3432
+ "loss": 0.0217,
3433
+ "step": 4890
3434
+ },
3435
+ {
3436
+ "epoch": 1.9514137793707687,
3437
+ "grad_norm": 0.34016168117523193,
3438
+ "learning_rate": 5.579065560136467e-05,
3439
+ "loss": 0.0253,
3440
+ "step": 4900
3441
+ },
3442
+ {
3443
+ "epoch": 1.9553962564715253,
3444
+ "grad_norm": 0.47430235147476196,
3445
+ "learning_rate": 5.562638987230392e-05,
3446
+ "loss": 0.0239,
3447
+ "step": 4910
3448
+ },
3449
+ {
3450
+ "epoch": 1.959378733572282,
3451
+ "grad_norm": 0.40641021728515625,
3452
+ "learning_rate": 5.546206261394498e-05,
3453
+ "loss": 0.0216,
3454
+ "step": 4920
3455
+ },
3456
+ {
3457
+ "epoch": 1.9633612106730385,
3458
+ "grad_norm": 0.39244401454925537,
3459
+ "learning_rate": 5.529767562334437e-05,
3460
+ "loss": 0.0236,
3461
+ "step": 4930
3462
+ },
3463
+ {
3464
+ "epoch": 1.9673436877737953,
3465
+ "grad_norm": 0.3444770276546478,
3466
+ "learning_rate": 5.5133230698211926e-05,
3467
+ "loss": 0.0229,
3468
+ "step": 4940
3469
+ },
3470
+ {
3471
+ "epoch": 1.971326164874552,
3472
+ "grad_norm": 0.40012794733047485,
3473
+ "learning_rate": 5.496872963689096e-05,
3474
+ "loss": 0.0233,
3475
+ "step": 4950
3476
+ },
3477
+ {
3478
+ "epoch": 1.9753086419753085,
3479
+ "grad_norm": 0.39745375514030457,
3480
+ "learning_rate": 5.4804174238338756e-05,
3481
+ "loss": 0.0205,
3482
+ "step": 4960
3483
+ },
3484
+ {
3485
+ "epoch": 1.9792911190760654,
3486
+ "grad_norm": 0.36429905891418457,
3487
+ "learning_rate": 5.463956630210678e-05,
3488
+ "loss": 0.0272,
3489
+ "step": 4970
3490
+ },
3491
+ {
3492
+ "epoch": 1.983273596176822,
3493
+ "grad_norm": 0.3775866627693176,
3494
+ "learning_rate": 5.4474907628321046e-05,
3495
+ "loss": 0.0254,
3496
+ "step": 4980
3497
+ },
3498
+ {
3499
+ "epoch": 1.9872560732775786,
3500
+ "grad_norm": 0.2797946333885193,
3501
+ "learning_rate": 5.431020001766244e-05,
3502
+ "loss": 0.0232,
3503
+ "step": 4990
3504
+ },
3505
+ {
3506
+ "epoch": 1.9912385503783354,
3507
+ "grad_norm": 0.31640321016311646,
3508
+ "learning_rate": 5.4145445271346986e-05,
3509
+ "loss": 0.0236,
3510
+ "step": 5000
3511
+ }
3512
+ ],
3513
+ "logging_steps": 10,
3514
+ "max_steps": 10000,
3515
+ "num_input_tokens_seen": 0,
3516
+ "num_train_epochs": 4,
3517
+ "save_steps": 1000,
3518
+ "stateful_callbacks": {
3519
+ "TrainerControl": {
3520
+ "args": {
3521
+ "should_epoch_stop": false,
3522
+ "should_evaluate": false,
3523
+ "should_log": false,
3524
+ "should_save": true,
3525
+ "should_training_stop": false
3526
+ },
3527
+ "attributes": {}
3528
+ }
3529
+ },
3530
+ "total_flos": 0.0,
3531
+ "train_batch_size": 32,
3532
+ "trial_name": null,
3533
+ "trial_params": null
3534
+ }
checkpoint-6000/config.json ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "action_dim": 32,
3
+ "action_head_cfg": {
4
+ "action_dim": 32,
5
+ "action_horizon": 16,
6
+ "add_pos_embed": true,
7
+ "backbone_embedding_dim": 2048,
8
+ "diffusion_model_cfg": {
9
+ "attention_head_dim": 48,
10
+ "cross_attention_dim": 2048,
11
+ "dropout": 0.2,
12
+ "final_dropout": true,
13
+ "interleave_self_attention": true,
14
+ "norm_type": "ada_norm",
15
+ "num_attention_heads": 32,
16
+ "num_layers": 16,
17
+ "output_dim": 1024,
18
+ "positional_embeddings": null
19
+ },
20
+ "hidden_size": 1024,
21
+ "input_embedding_dim": 1536,
22
+ "max_action_dim": 32,
23
+ "max_state_dim": 64,
24
+ "model_dtype": "float32",
25
+ "noise_beta_alpha": 1.5,
26
+ "noise_beta_beta": 1.0,
27
+ "noise_s": 0.999,
28
+ "num_inference_timesteps": 4,
29
+ "num_target_vision_tokens": 32,
30
+ "num_timestep_buckets": 1000,
31
+ "tune_diffusion_model": true,
32
+ "tune_projector": true,
33
+ "use_vlln": true,
34
+ "vl_self_attention_cfg": {
35
+ "attention_head_dim": 64,
36
+ "dropout": 0.2,
37
+ "final_dropout": true,
38
+ "num_attention_heads": 32,
39
+ "num_layers": 4,
40
+ "positional_embeddings": null
41
+ }
42
+ },
43
+ "action_horizon": 16,
44
+ "architectures": [
45
+ "GR00T_N1_5"
46
+ ],
47
+ "attn_implementation": null,
48
+ "backbone_cfg": {
49
+ "eagle_path": "NVEagle/eagle_er-qwen3_1_7B-Siglip2_400M_stage1_5_128gpu_er_v7_1mlp_nops",
50
+ "load_bf16": false,
51
+ "project_to_dim": null,
52
+ "reproject_vision": false,
53
+ "select_layer": 12,
54
+ "tune_llm": false,
55
+ "tune_visual": true,
56
+ "use_flash_attention": true
57
+ },
58
+ "compute_dtype": "bfloat16",
59
+ "hidden_size": 2048,
60
+ "model_dtype": "float32",
61
+ "model_type": "gr00t_n1_5",
62
+ "torch_dtype": "bfloat16",
63
+ "transformers_version": "4.51.3"
64
+ }
checkpoint-6000/experiment_cfg/metadata.json ADDED
@@ -0,0 +1,195 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "new_embodiment": {
3
+ "statistics": {
4
+ "state": {
5
+ "single_arm": {
6
+ "max": [
7
+ 33.22265625,
8
+ 86.1328125,
9
+ 84.7265625,
10
+ 101.337890625,
11
+ 32.783203125
12
+ ],
13
+ "min": [
14
+ -87.275390625,
15
+ -103.974609375,
16
+ -106.435546875,
17
+ -16.34765625,
18
+ -170.068359375
19
+ ],
20
+ "mean": [
21
+ -26.84341812133789,
22
+ 6.497307777404785,
23
+ -22.14896583557129,
24
+ 74.85757446289062,
25
+ -111.02351379394531
26
+ ],
27
+ "std": [
28
+ 27.50257110595703,
29
+ 41.151065826416016,
30
+ 48.41417694091797,
31
+ 20.208982467651367,
32
+ 46.13625717163086
33
+ ],
34
+ "q01": [
35
+ -79.013671875,
36
+ -103.271484375,
37
+ -104.0625,
38
+ 19.187402343750005,
39
+ -169.892578125
40
+ ],
41
+ "q99": [
42
+ 24.169921875,
43
+ 73.30078125,
44
+ 84.375,
45
+ 101.07421875,
46
+ 11.162109375
47
+ ]
48
+ },
49
+ "gripper": {
50
+ "max": [
51
+ 92.373046875
52
+ ],
53
+ "min": [
54
+ -0.615234375
55
+ ],
56
+ "mean": [
57
+ 14.521344184875488
58
+ ],
59
+ "std": [
60
+ 13.066176414489746
61
+ ],
62
+ "q01": [
63
+ 0.0
64
+ ],
65
+ "q99": [
66
+ 52.470703125
67
+ ]
68
+ }
69
+ },
70
+ "action": {
71
+ "single_arm": {
72
+ "max": [
73
+ 33.662109375,
74
+ 86.484375,
75
+ 85.078125,
76
+ 102.65625,
77
+ 33.3984375
78
+ ],
79
+ "min": [
80
+ -87.890625,
81
+ -104.4140625,
82
+ -110.7421875,
83
+ -19.599609375,
84
+ -173.3203125
85
+ ],
86
+ "mean": [
87
+ -26.84657096862793,
88
+ 4.594353199005127,
89
+ -23.747364044189453,
90
+ 74.75859832763672,
91
+ -111.30194854736328
92
+ ],
93
+ "std": [
94
+ 27.627857208251953,
95
+ 40.672019958496094,
96
+ 48.966896057128906,
97
+ 20.728727340698242,
98
+ 46.57689666748047
99
+ ],
100
+ "q01": [
101
+ -79.453125,
102
+ -102.744140625,
103
+ -109.6875,
104
+ 17.9296875,
105
+ -172.177734375
106
+ ],
107
+ "q99": [
108
+ 24.08203125,
109
+ 72.685546875,
110
+ 83.84765625,
111
+ 101.77734375,
112
+ 10.72265625
113
+ ]
114
+ },
115
+ "gripper": {
116
+ "max": [
117
+ 92.63671875
118
+ ],
119
+ "min": [
120
+ -3.076171875
121
+ ],
122
+ "mean": [
123
+ 12.46182632446289
124
+ ],
125
+ "std": [
126
+ 14.029601097106934
127
+ ],
128
+ "q01": [
129
+ -0.52734375
130
+ ],
131
+ "q99": [
132
+ 52.91015625
133
+ ]
134
+ }
135
+ }
136
+ },
137
+ "modalities": {
138
+ "video": {
139
+ "front": {
140
+ "resolution": [
141
+ 640,
142
+ 480
143
+ ],
144
+ "channels": 3,
145
+ "fps": 30.0
146
+ },
147
+ "laptop": {
148
+ "resolution": [
149
+ 640,
150
+ 480
151
+ ],
152
+ "channels": 3,
153
+ "fps": 30.0
154
+ }
155
+ },
156
+ "state": {
157
+ "single_arm": {
158
+ "absolute": true,
159
+ "rotation_type": null,
160
+ "shape": [
161
+ 5
162
+ ],
163
+ "continuous": true
164
+ },
165
+ "gripper": {
166
+ "absolute": true,
167
+ "rotation_type": null,
168
+ "shape": [
169
+ 1
170
+ ],
171
+ "continuous": true
172
+ }
173
+ },
174
+ "action": {
175
+ "single_arm": {
176
+ "absolute": true,
177
+ "rotation_type": null,
178
+ "shape": [
179
+ 5
180
+ ],
181
+ "continuous": true
182
+ },
183
+ "gripper": {
184
+ "absolute": true,
185
+ "rotation_type": null,
186
+ "shape": [
187
+ 1
188
+ ],
189
+ "continuous": true
190
+ }
191
+ }
192
+ },
193
+ "embodiment_tag": "new_embodiment"
194
+ }
195
+ }
checkpoint-6000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0e65895a7b6c052e58b679bdd68bff54a68694a81f38e23d3e03a92059b3600b
3
+ size 4999367032
checkpoint-6000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:185c359b4bba0fb8e8f070cc9219b2dc7ef2b9222aa348ef4d29659957794091
3
+ size 2586705312
checkpoint-6000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-6000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:145c4fbd2b48ae7a568038fa1ecc19b796c1fce35b0883b3c0ce72c3a956a2a6
3
+ size 4147424698
checkpoint-6000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:13bd3aee67d899586785f110c683415e1f00044b0294c797f6849a5b51cc801a
3
+ size 14244
checkpoint-6000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8d6c0bf646ac78d188a2258804081d2e369772e8091e5395d9073771ec310aca
3
+ size 1064
checkpoint-6000/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-7000/config.json ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "action_dim": 32,
3
+ "action_head_cfg": {
4
+ "action_dim": 32,
5
+ "action_horizon": 16,
6
+ "add_pos_embed": true,
7
+ "backbone_embedding_dim": 2048,
8
+ "diffusion_model_cfg": {
9
+ "attention_head_dim": 48,
10
+ "cross_attention_dim": 2048,
11
+ "dropout": 0.2,
12
+ "final_dropout": true,
13
+ "interleave_self_attention": true,
14
+ "norm_type": "ada_norm",
15
+ "num_attention_heads": 32,
16
+ "num_layers": 16,
17
+ "output_dim": 1024,
18
+ "positional_embeddings": null
19
+ },
20
+ "hidden_size": 1024,
21
+ "input_embedding_dim": 1536,
22
+ "max_action_dim": 32,
23
+ "max_state_dim": 64,
24
+ "model_dtype": "float32",
25
+ "noise_beta_alpha": 1.5,
26
+ "noise_beta_beta": 1.0,
27
+ "noise_s": 0.999,
28
+ "num_inference_timesteps": 4,
29
+ "num_target_vision_tokens": 32,
30
+ "num_timestep_buckets": 1000,
31
+ "tune_diffusion_model": true,
32
+ "tune_projector": true,
33
+ "use_vlln": true,
34
+ "vl_self_attention_cfg": {
35
+ "attention_head_dim": 64,
36
+ "dropout": 0.2,
37
+ "final_dropout": true,
38
+ "num_attention_heads": 32,
39
+ "num_layers": 4,
40
+ "positional_embeddings": null
41
+ }
42
+ },
43
+ "action_horizon": 16,
44
+ "architectures": [
45
+ "GR00T_N1_5"
46
+ ],
47
+ "attn_implementation": null,
48
+ "backbone_cfg": {
49
+ "eagle_path": "NVEagle/eagle_er-qwen3_1_7B-Siglip2_400M_stage1_5_128gpu_er_v7_1mlp_nops",
50
+ "load_bf16": false,
51
+ "project_to_dim": null,
52
+ "reproject_vision": false,
53
+ "select_layer": 12,
54
+ "tune_llm": false,
55
+ "tune_visual": true,
56
+ "use_flash_attention": true
57
+ },
58
+ "compute_dtype": "bfloat16",
59
+ "hidden_size": 2048,
60
+ "model_dtype": "float32",
61
+ "model_type": "gr00t_n1_5",
62
+ "torch_dtype": "bfloat16",
63
+ "transformers_version": "4.51.3"
64
+ }
checkpoint-7000/experiment_cfg/metadata.json ADDED
@@ -0,0 +1,195 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "new_embodiment": {
3
+ "statistics": {
4
+ "state": {
5
+ "single_arm": {
6
+ "max": [
7
+ 33.22265625,
8
+ 86.1328125,
9
+ 84.7265625,
10
+ 101.337890625,
11
+ 32.783203125
12
+ ],
13
+ "min": [
14
+ -87.275390625,
15
+ -103.974609375,
16
+ -106.435546875,
17
+ -16.34765625,
18
+ -170.068359375
19
+ ],
20
+ "mean": [
21
+ -26.84341812133789,
22
+ 6.497307777404785,
23
+ -22.14896583557129,
24
+ 74.85757446289062,
25
+ -111.02351379394531
26
+ ],
27
+ "std": [
28
+ 27.50257110595703,
29
+ 41.151065826416016,
30
+ 48.41417694091797,
31
+ 20.208982467651367,
32
+ 46.13625717163086
33
+ ],
34
+ "q01": [
35
+ -79.013671875,
36
+ -103.271484375,
37
+ -104.0625,
38
+ 19.187402343750005,
39
+ -169.892578125
40
+ ],
41
+ "q99": [
42
+ 24.169921875,
43
+ 73.30078125,
44
+ 84.375,
45
+ 101.07421875,
46
+ 11.162109375
47
+ ]
48
+ },
49
+ "gripper": {
50
+ "max": [
51
+ 92.373046875
52
+ ],
53
+ "min": [
54
+ -0.615234375
55
+ ],
56
+ "mean": [
57
+ 14.521344184875488
58
+ ],
59
+ "std": [
60
+ 13.066176414489746
61
+ ],
62
+ "q01": [
63
+ 0.0
64
+ ],
65
+ "q99": [
66
+ 52.470703125
67
+ ]
68
+ }
69
+ },
70
+ "action": {
71
+ "single_arm": {
72
+ "max": [
73
+ 33.662109375,
74
+ 86.484375,
75
+ 85.078125,
76
+ 102.65625,
77
+ 33.3984375
78
+ ],
79
+ "min": [
80
+ -87.890625,
81
+ -104.4140625,
82
+ -110.7421875,
83
+ -19.599609375,
84
+ -173.3203125
85
+ ],
86
+ "mean": [
87
+ -26.84657096862793,
88
+ 4.594353199005127,
89
+ -23.747364044189453,
90
+ 74.75859832763672,
91
+ -111.30194854736328
92
+ ],
93
+ "std": [
94
+ 27.627857208251953,
95
+ 40.672019958496094,
96
+ 48.966896057128906,
97
+ 20.728727340698242,
98
+ 46.57689666748047
99
+ ],
100
+ "q01": [
101
+ -79.453125,
102
+ -102.744140625,
103
+ -109.6875,
104
+ 17.9296875,
105
+ -172.177734375
106
+ ],
107
+ "q99": [
108
+ 24.08203125,
109
+ 72.685546875,
110
+ 83.84765625,
111
+ 101.77734375,
112
+ 10.72265625
113
+ ]
114
+ },
115
+ "gripper": {
116
+ "max": [
117
+ 92.63671875
118
+ ],
119
+ "min": [
120
+ -3.076171875
121
+ ],
122
+ "mean": [
123
+ 12.46182632446289
124
+ ],
125
+ "std": [
126
+ 14.029601097106934
127
+ ],
128
+ "q01": [
129
+ -0.52734375
130
+ ],
131
+ "q99": [
132
+ 52.91015625
133
+ ]
134
+ }
135
+ }
136
+ },
137
+ "modalities": {
138
+ "video": {
139
+ "front": {
140
+ "resolution": [
141
+ 640,
142
+ 480
143
+ ],
144
+ "channels": 3,
145
+ "fps": 30.0
146
+ },
147
+ "laptop": {
148
+ "resolution": [
149
+ 640,
150
+ 480
151
+ ],
152
+ "channels": 3,
153
+ "fps": 30.0
154
+ }
155
+ },
156
+ "state": {
157
+ "single_arm": {
158
+ "absolute": true,
159
+ "rotation_type": null,
160
+ "shape": [
161
+ 5
162
+ ],
163
+ "continuous": true
164
+ },
165
+ "gripper": {
166
+ "absolute": true,
167
+ "rotation_type": null,
168
+ "shape": [
169
+ 1
170
+ ],
171
+ "continuous": true
172
+ }
173
+ },
174
+ "action": {
175
+ "single_arm": {
176
+ "absolute": true,
177
+ "rotation_type": null,
178
+ "shape": [
179
+ 5
180
+ ],
181
+ "continuous": true
182
+ },
183
+ "gripper": {
184
+ "absolute": true,
185
+ "rotation_type": null,
186
+ "shape": [
187
+ 1
188
+ ],
189
+ "continuous": true
190
+ }
191
+ }
192
+ },
193
+ "embodiment_tag": "new_embodiment"
194
+ }
195
+ }
checkpoint-7000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0e65895a7b6c052e58b679bdd68bff54a68694a81f38e23d3e03a92059b3600b
3
+ size 4999367032
checkpoint-7000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:34e7230b2f52e65b83629e430d28e02bffb786970b756c13cd4a21211b800f5a
3
+ size 2586705312
checkpoint-7000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-7000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:14bc9ed5dcc9ab13c182c3a8d5c25bc5f9d6e9cf1efae67134d2e683a75c39e1
3
+ size 4147424698
checkpoint-7000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ebdbf47e1b77224d159715275bc578926a4cd08f83b15b47e0225eb3eab56e59
3
+ size 14244
checkpoint-7000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:45049460c0fa55fae34b1884c511d0c5cb5241d0d636fad92c82b0e0a657580e
3
+ size 1064
checkpoint-7000/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-8000/config.json ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "action_dim": 32,
3
+ "action_head_cfg": {
4
+ "action_dim": 32,
5
+ "action_horizon": 16,
6
+ "add_pos_embed": true,
7
+ "backbone_embedding_dim": 2048,
8
+ "diffusion_model_cfg": {
9
+ "attention_head_dim": 48,
10
+ "cross_attention_dim": 2048,
11
+ "dropout": 0.2,
12
+ "final_dropout": true,
13
+ "interleave_self_attention": true,
14
+ "norm_type": "ada_norm",
15
+ "num_attention_heads": 32,
16
+ "num_layers": 16,
17
+ "output_dim": 1024,
18
+ "positional_embeddings": null
19
+ },
20
+ "hidden_size": 1024,
21
+ "input_embedding_dim": 1536,
22
+ "max_action_dim": 32,
23
+ "max_state_dim": 64,
24
+ "model_dtype": "float32",
25
+ "noise_beta_alpha": 1.5,
26
+ "noise_beta_beta": 1.0,
27
+ "noise_s": 0.999,
28
+ "num_inference_timesteps": 4,
29
+ "num_target_vision_tokens": 32,
30
+ "num_timestep_buckets": 1000,
31
+ "tune_diffusion_model": true,
32
+ "tune_projector": true,
33
+ "use_vlln": true,
34
+ "vl_self_attention_cfg": {
35
+ "attention_head_dim": 64,
36
+ "dropout": 0.2,
37
+ "final_dropout": true,
38
+ "num_attention_heads": 32,
39
+ "num_layers": 4,
40
+ "positional_embeddings": null
41
+ }
42
+ },
43
+ "action_horizon": 16,
44
+ "architectures": [
45
+ "GR00T_N1_5"
46
+ ],
47
+ "attn_implementation": null,
48
+ "backbone_cfg": {
49
+ "eagle_path": "NVEagle/eagle_er-qwen3_1_7B-Siglip2_400M_stage1_5_128gpu_er_v7_1mlp_nops",
50
+ "load_bf16": false,
51
+ "project_to_dim": null,
52
+ "reproject_vision": false,
53
+ "select_layer": 12,
54
+ "tune_llm": false,
55
+ "tune_visual": true,
56
+ "use_flash_attention": true
57
+ },
58
+ "compute_dtype": "bfloat16",
59
+ "hidden_size": 2048,
60
+ "model_dtype": "float32",
61
+ "model_type": "gr00t_n1_5",
62
+ "torch_dtype": "bfloat16",
63
+ "transformers_version": "4.51.3"
64
+ }
checkpoint-8000/experiment_cfg/metadata.json ADDED
@@ -0,0 +1,195 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "new_embodiment": {
3
+ "statistics": {
4
+ "state": {
5
+ "single_arm": {
6
+ "max": [
7
+ 33.22265625,
8
+ 86.1328125,
9
+ 84.7265625,
10
+ 101.337890625,
11
+ 32.783203125
12
+ ],
13
+ "min": [
14
+ -87.275390625,
15
+ -103.974609375,
16
+ -106.435546875,
17
+ -16.34765625,
18
+ -170.068359375
19
+ ],
20
+ "mean": [
21
+ -26.84341812133789,
22
+ 6.497307777404785,
23
+ -22.14896583557129,
24
+ 74.85757446289062,
25
+ -111.02351379394531
26
+ ],
27
+ "std": [
28
+ 27.50257110595703,
29
+ 41.151065826416016,
30
+ 48.41417694091797,
31
+ 20.208982467651367,
32
+ 46.13625717163086
33
+ ],
34
+ "q01": [
35
+ -79.013671875,
36
+ -103.271484375,
37
+ -104.0625,
38
+ 19.187402343750005,
39
+ -169.892578125
40
+ ],
41
+ "q99": [
42
+ 24.169921875,
43
+ 73.30078125,
44
+ 84.375,
45
+ 101.07421875,
46
+ 11.162109375
47
+ ]
48
+ },
49
+ "gripper": {
50
+ "max": [
51
+ 92.373046875
52
+ ],
53
+ "min": [
54
+ -0.615234375
55
+ ],
56
+ "mean": [
57
+ 14.521344184875488
58
+ ],
59
+ "std": [
60
+ 13.066176414489746
61
+ ],
62
+ "q01": [
63
+ 0.0
64
+ ],
65
+ "q99": [
66
+ 52.470703125
67
+ ]
68
+ }
69
+ },
70
+ "action": {
71
+ "single_arm": {
72
+ "max": [
73
+ 33.662109375,
74
+ 86.484375,
75
+ 85.078125,
76
+ 102.65625,
77
+ 33.3984375
78
+ ],
79
+ "min": [
80
+ -87.890625,
81
+ -104.4140625,
82
+ -110.7421875,
83
+ -19.599609375,
84
+ -173.3203125
85
+ ],
86
+ "mean": [
87
+ -26.84657096862793,
88
+ 4.594353199005127,
89
+ -23.747364044189453,
90
+ 74.75859832763672,
91
+ -111.30194854736328
92
+ ],
93
+ "std": [
94
+ 27.627857208251953,
95
+ 40.672019958496094,
96
+ 48.966896057128906,
97
+ 20.728727340698242,
98
+ 46.57689666748047
99
+ ],
100
+ "q01": [
101
+ -79.453125,
102
+ -102.744140625,
103
+ -109.6875,
104
+ 17.9296875,
105
+ -172.177734375
106
+ ],
107
+ "q99": [
108
+ 24.08203125,
109
+ 72.685546875,
110
+ 83.84765625,
111
+ 101.77734375,
112
+ 10.72265625
113
+ ]
114
+ },
115
+ "gripper": {
116
+ "max": [
117
+ 92.63671875
118
+ ],
119
+ "min": [
120
+ -3.076171875
121
+ ],
122
+ "mean": [
123
+ 12.46182632446289
124
+ ],
125
+ "std": [
126
+ 14.029601097106934
127
+ ],
128
+ "q01": [
129
+ -0.52734375
130
+ ],
131
+ "q99": [
132
+ 52.91015625
133
+ ]
134
+ }
135
+ }
136
+ },
137
+ "modalities": {
138
+ "video": {
139
+ "front": {
140
+ "resolution": [
141
+ 640,
142
+ 480
143
+ ],
144
+ "channels": 3,
145
+ "fps": 30.0
146
+ },
147
+ "laptop": {
148
+ "resolution": [
149
+ 640,
150
+ 480
151
+ ],
152
+ "channels": 3,
153
+ "fps": 30.0
154
+ }
155
+ },
156
+ "state": {
157
+ "single_arm": {
158
+ "absolute": true,
159
+ "rotation_type": null,
160
+ "shape": [
161
+ 5
162
+ ],
163
+ "continuous": true
164
+ },
165
+ "gripper": {
166
+ "absolute": true,
167
+ "rotation_type": null,
168
+ "shape": [
169
+ 1
170
+ ],
171
+ "continuous": true
172
+ }
173
+ },
174
+ "action": {
175
+ "single_arm": {
176
+ "absolute": true,
177
+ "rotation_type": null,
178
+ "shape": [
179
+ 5
180
+ ],
181
+ "continuous": true
182
+ },
183
+ "gripper": {
184
+ "absolute": true,
185
+ "rotation_type": null,
186
+ "shape": [
187
+ 1
188
+ ],
189
+ "continuous": true
190
+ }
191
+ }
192
+ },
193
+ "embodiment_tag": "new_embodiment"
194
+ }
195
+ }
checkpoint-8000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0e65895a7b6c052e58b679bdd68bff54a68694a81f38e23d3e03a92059b3600b
3
+ size 4999367032
checkpoint-8000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:55b8a95b9a98c61059711fe605141c966d8029578f3817615ff85066850b7fa0
3
+ size 2586705312
checkpoint-8000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff