badaoui HF Staff commited on
Commit
7eec7dd
·
verified ·
1 Parent(s): 8479ec5

Synchronizing local compiler cache (#7)

Browse files

- Synchronizing local compiler cache (7f22736ee09cff8168c8692ba49709b2ab056317)

.gitattributes CHANGED
@@ -59,3 +59,7 @@ neuronxcc-2.19.8089.0+8ab9f450/MODULE_77e03274751ff8d8f0ef/text_encoder/model.ne
59
  neuronxcc-2.19.8089.0+8ab9f450/MODULE_77e03274751ff8d8f0ef/transformer/model.neuron filter=lfs diff=lfs merge=lfs -text
60
  neuronxcc-2.19.8089.0+8ab9f450/MODULE_77e03274751ff8d8f0ef/vae_decoder/model.neuron filter=lfs diff=lfs merge=lfs -text
61
  neuronxcc-2.19.8089.0+8ab9f450/MODULE_77e03274751ff8d8f0ef/vae_encoder/model.neuron filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
59
  neuronxcc-2.19.8089.0+8ab9f450/MODULE_77e03274751ff8d8f0ef/transformer/model.neuron filter=lfs diff=lfs merge=lfs -text
60
  neuronxcc-2.19.8089.0+8ab9f450/MODULE_77e03274751ff8d8f0ef/vae_decoder/model.neuron filter=lfs diff=lfs merge=lfs -text
61
  neuronxcc-2.19.8089.0+8ab9f450/MODULE_77e03274751ff8d8f0ef/vae_encoder/model.neuron filter=lfs diff=lfs merge=lfs -text
62
+ neuronxcc-2.19.8089.0+8ab9f450/MODULE_6dc18b51a0bd41e6db9b+ed72d204/model.neff filter=lfs diff=lfs merge=lfs -text
63
+ neuronxcc-2.19.8089.0+8ab9f450/MODULE_bbfac8bf25e0db852ad6+431f5505/model.neff filter=lfs diff=lfs merge=lfs -text
64
+ neuronxcc-2.19.8089.0+8ab9f450/MODULE_c6adfb0a32e0702731a6+a9d440f5/model.neff filter=lfs diff=lfs merge=lfs -text
65
+ neuronxcc-2.19.8089.0+8ab9f450/MODULE_c6adfb0a32e0702731a6+a9d440f5/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
neuronxcc-2.19.8089.0+8ab9f450/0_REGISTRY/0.3.0/phi3/microsoft/Phi-3-mini-128k-instruct/03724e93788dd5ca4d4a.json ADDED
@@ -0,0 +1,179 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_entry_class": "SingleModelCacheEntry",
3
+ "_model_id": "microsoft/Phi-3-mini-128k-instruct",
4
+ "_task": "text-generation",
5
+ "architectures": [
6
+ "Phi3ForCausalLM"
7
+ ],
8
+ "attention_bias": false,
9
+ "attention_dropout": 0.0,
10
+ "auto_map": {
11
+ "AutoConfig": "microsoft/Phi-3-mini-128k-instruct--configuration_phi3.Phi3Config",
12
+ "AutoModelForCausalLM": "microsoft/Phi-3-mini-128k-instruct--modeling_phi3.Phi3ForCausalLM"
13
+ },
14
+ "embd_pdrop": 0.0,
15
+ "hidden_act": "silu",
16
+ "hidden_size": 3072,
17
+ "initializer_range": 0.02,
18
+ "intermediate_size": 8192,
19
+ "max_position_embeddings": 131072,
20
+ "model_type": "phi3",
21
+ "neuron": {
22
+ "_serialized_key": "NxDNeuronConfig",
23
+ "async_mode": false,
24
+ "attn_kernel_enabled": false,
25
+ "batch_size": 1,
26
+ "capacity_factor": null,
27
+ "cc_pipeline_tiling_factor": 2,
28
+ "checkpoint_id": "microsoft/Phi-3-mini-128k-instruct",
29
+ "checkpoint_revision": "072cb7562cb8c4adf682a8e186aaafa49469eb5d",
30
+ "continuous_batching": false,
31
+ "enable_bucketing": false,
32
+ "ep_degree": 1,
33
+ "flash_decoding_enabled": false,
34
+ "fused_qkv": true,
35
+ "glu_mlp": true,
36
+ "is_chunked_prefill": false,
37
+ "local_ranks_size": 12,
38
+ "logical_nc_config": 1,
39
+ "max_batch_size": 1,
40
+ "max_context_length": 128,
41
+ "max_topk": 256,
42
+ "mlp_kernel_enabled": false,
43
+ "mlp_kernel_fuse_residual_add": false,
44
+ "n_active_tokens": 128,
45
+ "neuronxcc_version": "2.19.8089.0+8ab9f450",
46
+ "num_cores_per_group": 1,
47
+ "on_device_sampling": true,
48
+ "optimum_neuron_version": "0.3.0",
49
+ "output_logits": false,
50
+ "padding_side": "right",
51
+ "pp_degree": 1,
52
+ "qk_layernorm": false,
53
+ "qkv_kernel_enabled": false,
54
+ "rpl_reduce_dtype": "bfloat16",
55
+ "sequence_length": 128,
56
+ "sequence_parallel_enabled": false,
57
+ "speculation_length": 0,
58
+ "start_rank_id": 0,
59
+ "target": null,
60
+ "torch_dtype": "bfloat16",
61
+ "tp_degree": 12,
62
+ "vocab_parallel": false
63
+ },
64
+ "num_attention_heads": 32,
65
+ "num_hidden_layers": 32,
66
+ "num_key_value_heads": 32,
67
+ "original_max_position_embeddings": 4096,
68
+ "partial_rotary_factor": 1.0,
69
+ "resid_pdrop": 0.0,
70
+ "rms_norm_eps": 1e-05,
71
+ "rope_scaling": {
72
+ "long_factor": [
73
+ 1.0700000524520874,
74
+ 1.1200000047683716,
75
+ 1.149999976158142,
76
+ 1.4199999570846558,
77
+ 1.5699999332427979,
78
+ 1.7999999523162842,
79
+ 2.129999876022339,
80
+ 2.129999876022339,
81
+ 3.009999990463257,
82
+ 5.910000324249268,
83
+ 6.950000286102295,
84
+ 9.070000648498535,
85
+ 9.930000305175781,
86
+ 10.710000038146973,
87
+ 11.130000114440918,
88
+ 14.609999656677246,
89
+ 15.409998893737793,
90
+ 19.809999465942383,
91
+ 37.279998779296875,
92
+ 38.279998779296875,
93
+ 38.599998474121094,
94
+ 40.12000274658203,
95
+ 46.20000457763672,
96
+ 50.940006256103516,
97
+ 53.66000747680664,
98
+ 54.9373893737793,
99
+ 56.89738845825195,
100
+ 57.28738784790039,
101
+ 59.98738479614258,
102
+ 60.86738586425781,
103
+ 60.887386322021484,
104
+ 61.71739196777344,
105
+ 62.91739273071289,
106
+ 62.957393646240234,
107
+ 63.41739273071289,
108
+ 63.8173942565918,
109
+ 63.83739471435547,
110
+ 63.897396087646484,
111
+ 63.93739700317383,
112
+ 64.06739807128906,
113
+ 64.11434936523438,
114
+ 64.12435150146484,
115
+ 64.15435028076172,
116
+ 64.19435119628906,
117
+ 64.24435424804688,
118
+ 64.57435607910156,
119
+ 64.69000244140625,
120
+ 64.76000213623047
121
+ ],
122
+ "short_factor": [
123
+ 1.1,
124
+ 1.1,
125
+ 1.1,
126
+ 1.3000000000000003,
127
+ 1.3500000000000003,
128
+ 1.3500000000000003,
129
+ 1.4000000000000004,
130
+ 1.5500000000000005,
131
+ 2.000000000000001,
132
+ 2.000000000000001,
133
+ 2.000000000000001,
134
+ 2.000000000000001,
135
+ 2.000000000000001,
136
+ 2.000000000000001,
137
+ 2.000000000000001,
138
+ 2.000000000000001,
139
+ 2.000000000000001,
140
+ 2.000000000000001,
141
+ 2.000000000000001,
142
+ 2.000000000000001,
143
+ 2.000000000000001,
144
+ 2.000000000000001,
145
+ 2.000000000000001,
146
+ 2.000000000000001,
147
+ 2.000000000000001,
148
+ 2.0500000000000007,
149
+ 2.0500000000000007,
150
+ 2.0500000000000007,
151
+ 2.0500000000000007,
152
+ 2.0500000000000007,
153
+ 2.0500000000000007,
154
+ 2.1000000000000005,
155
+ 2.1000000000000005,
156
+ 2.1500000000000004,
157
+ 2.25,
158
+ 2.25,
159
+ 2.25,
160
+ 2.25,
161
+ 2.25,
162
+ 2.3999999999999995,
163
+ 2.4499999999999993,
164
+ 2.499999999999999,
165
+ 2.6999999999999984,
166
+ 2.6999999999999984,
167
+ 2.7499999999999982,
168
+ 2.799999999999998,
169
+ 2.8999999999999977,
170
+ 3.049999999999997
171
+ ],
172
+ "type": "longrope"
173
+ },
174
+ "rope_theta": 10000.0,
175
+ "sliding_window": 262144,
176
+ "tie_word_embeddings": false,
177
+ "use_cache": true,
178
+ "vocab_size": 32064
179
+ }
neuronxcc-2.19.8089.0+8ab9f450/MODULE_6dc18b51a0bd41e6db9b+ed72d204/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ ["--target=trn1", "--auto-cast=none", "--model-type=transformer", "--tensorizer-options=--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-strided-dma ", "-O2", "--lnc=1", "--logfile=/tmp/nxd_model/context_encoding_model/_tp0_bk0/log-neuron-cc.txt"]
neuronxcc-2.19.8089.0+8ab9f450/MODULE_6dc18b51a0bd41e6db9b+ed72d204/model.done ADDED
File without changes
neuronxcc-2.19.8089.0+8ab9f450/MODULE_6dc18b51a0bd41e6db9b+ed72d204/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:19faefd1acbb0f9b00f4227f76db71e5446ac3c75c3436f8b07aca41ad2187bf
3
+ size 644860
neuronxcc-2.19.8089.0+8ab9f450/MODULE_6dc18b51a0bd41e6db9b+ed72d204/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cf14bfa46177d9a98b6382e6e3641e1bf03de4fcb36f5a6a0880af0f3d827f28
3
+ size 973824
neuronxcc-2.19.8089.0+8ab9f450/MODULE_bbfac8bf25e0db852ad6+431f5505/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ "--model-type=transformer -O1 --lnc=1 --internal-hlo2tensorizer-options=--experimental-unsafe-fp8e4m3fn-as-fp8e4m3 --logfile=/tmp/nxd_model/layout_opt/log-neuron-cc.txt"
neuronxcc-2.19.8089.0+8ab9f450/MODULE_bbfac8bf25e0db852ad6+431f5505/model.done ADDED
File without changes
neuronxcc-2.19.8089.0+8ab9f450/MODULE_bbfac8bf25e0db852ad6+431f5505/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b2cb04dc262780273d065f9446e567b6914c274dd8c966e5201383c4424b5abb
3
+ size 135302
neuronxcc-2.19.8089.0+8ab9f450/MODULE_bbfac8bf25e0db852ad6+431f5505/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:edf5765d498e20b9bbdee4ac7d819f3659a7d7eb5fcc759c969343604f828ac9
3
+ size 994304
neuronxcc-2.19.8089.0+8ab9f450/MODULE_c6adfb0a32e0702731a6+a9d440f5/compile_flags.json ADDED
@@ -0,0 +1 @@
 
 
1
+ ["--target=trn1", "--auto-cast=none", "--model-type=transformer", "--tensorizer-options=--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-strided-dma ", "-O2", "--lnc=1", "--logfile=/tmp/nxd_model/token_generation_model/_tp0_bk0/log-neuron-cc.txt", "--enable-internal-neff-wrapper"]
neuronxcc-2.19.8089.0+8ab9f450/MODULE_c6adfb0a32e0702731a6+a9d440f5/model.done ADDED
File without changes
neuronxcc-2.19.8089.0+8ab9f450/MODULE_c6adfb0a32e0702731a6+a9d440f5/model.hlo_module.pb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4688d6bc85f0ed616bc9a4efab0fe6d4296e04bab8bb16a1ff1988ebd5221182
3
+ size 696469
neuronxcc-2.19.8089.0+8ab9f450/MODULE_c6adfb0a32e0702731a6+a9d440f5/model.neff ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:13a1c17720c07ebef7155c31a549fd2282b3b4285a1fe0f526b9da4bdba114f3
3
+ size 1342464
neuronxcc-2.19.8089.0+8ab9f450/MODULE_c6adfb0a32e0702731a6+a9d440f5/wrapped_neff.hlo ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4e1507420ffcccbd61fed5642df510804ae7765785207a0d861cef82a174b4b2
3
+ size 1479315