Hugging Face's logo

yalhessi
/
lemexp-task1-v2-lemma_object_full-deepseek-coder-1.3b-base-ddp-8lr-v2

PEFT
Safetensors
Generated from Trainer
Model card Files
xet
Community
yalhessi commited on
Commit
67fff9c
·
verified ·
1 Parent(s): a8056a3

End of training

Browse files
Files changed (5) hide show
  1. README.md +64 -34
  2. adapter_config.json +2 -2
  3. adapter_model.safetensors +1 -1
  4. loss_plot.png +0 -0
  5. training_args.bin +2 -2
README.md CHANGED
@@ -16,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
16
 
17
  This model is a fine-tuned version of [deepseek-ai/deepseek-coder-1.3b-base](https://huggingface.co/deepseek-ai/deepseek-coder-1.3b-base) on an unknown dataset.
18
  It achieves the following results on the evaluation set:
19
- - Loss: 0.2570
20
 
21
  ## Model description
22
 
@@ -45,43 +45,73 @@ The following hyperparameters were used during training:
45
  - total_eval_batch_size: 16
46
  - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
47
  - lr_scheduler_type: linear
48
- - num_epochs: 6
49
  - mixed_precision_training: Native AMP
50
 
51
  ### Training results
52
 
53
- | Training Loss | Epoch | Step | Validation Loss |
54
- |:-------------:|:-----:|:-----:|:---------------:|
55
- | 0.4506 | 0.2 | 3094 | 0.4509 |
56
- | 0.4176 | 0.4 | 6188 | 0.4119 |
57
- | 0.4028 | 0.6 | 9282 | 0.4009 |
58
- | 0.3903 | 0.8 | 12376 | 0.3867 |
59
- | 0.3791 | 1.0 | 15470 | 0.3844 |
60
- | 0.3728 | 1.2 | 18564 | 0.3752 |
61
- | 0.3652 | 1.4 | 21658 | 0.3608 |
62
- | 0.3604 | 1.6 | 24752 | 0.3574 |
63
- | 0.3549 | 1.8 | 27846 | 0.3554 |
64
- | 0.3491 | 2.0 | 30940 | 0.3493 |
65
- | 0.3411 | 2.2 | 34034 | 0.3406 |
66
- | 0.3369 | 2.4 | 37128 | 0.3315 |
67
- | 0.3304 | 2.6 | 40222 | 0.3313 |
68
- | 0.3269 | 2.8 | 43316 | 0.3309 |
69
- | 0.3229 | 3.0 | 46410 | 0.3285 |
70
- | 0.3128 | 3.2 | 49504 | 0.3141 |
71
- | 0.3128 | 3.4 | 52598 | 0.3127 |
72
- | 0.3059 | 3.6 | 55692 | 0.3097 |
73
- | 0.3047 | 3.8 | 58786 | 0.3038 |
74
- | 0.3003 | 4.0 | 61880 | 0.2949 |
75
- | 0.2881 | 4.2 | 64974 | 0.2886 |
76
- | 0.2838 | 4.4 | 68068 | 0.2920 |
77
- | 0.2821 | 4.6 | 71162 | 0.2878 |
78
- | 0.2735 | 4.8 | 74256 | 0.2808 |
79
- | 0.2698 | 5.0 | 77350 | 0.2764 |
80
- | 0.2596 | 5.2 | 80444 | 0.2720 |
81
- | 0.2624 | 5.4 | 83538 | 0.2714 |
82
- | 0.2574 | 5.6 | 86632 | 0.2691 |
83
- | 0.2542 | 5.8 | 89726 | 0.2630 |
84
- | 0.2484 | 6.0 | 92820 | 0.2570 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
85
 
86
 
87
  ### Framework versions
 
16
 
17
  This model is a fine-tuned version of [deepseek-ai/deepseek-coder-1.3b-base](https://huggingface.co/deepseek-ai/deepseek-coder-1.3b-base) on an unknown dataset.
18
  It achieves the following results on the evaluation set:
19
+ - Loss: 0.2261
20
 
21
  ## Model description
22
 
 
45
  - total_eval_batch_size: 16
46
  - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
47
  - lr_scheduler_type: linear
48
+ - num_epochs: 12
49
  - mixed_precision_training: Native AMP
50
 
51
  ### Training results
52
 
53
+ | Training Loss | Epoch | Step | Validation Loss |
54
+ |:-------------:|:-----:|:------:|:---------------:|
55
+ | 0.4519 | 0.2 | 3094 | 0.4528 |
56
+ | 0.4192 | 0.4 | 6188 | 0.4162 |
57
+ | 0.4051 | 0.6 | 9282 | 0.4043 |
58
+ | 0.3928 | 0.8 | 12376 | 0.3904 |
59
+ | 0.3846 | 1.0 | 15470 | 0.3827 |
60
+ | 0.3797 | 1.2 | 18564 | 0.3772 |
61
+ | 0.3744 | 1.4 | 21658 | 0.3697 |
62
+ | 0.3697 | 1.6 | 24752 | 0.3640 |
63
+ | 0.3643 | 1.8 | 27846 | 0.3624 |
64
+ | 0.3614 | 2.0 | 30940 | 0.3526 |
65
+ | 0.3546 | 2.2 | 34034 | 0.3512 |
66
+ | 0.3503 | 2.4 | 37128 | 0.3487 |
67
+ | 0.345 | 2.6 | 40222 | 0.3421 |
68
+ | 0.3449 | 2.8 | 43316 | 0.3431 |
69
+ | 0.3421 | 3.0 | 46410 | 0.3432 |
70
+ | 0.335 | 3.2 | 49504 | 0.3359 |
71
+ | 0.3351 | 3.4 | 52598 | 0.3336 |
72
+ | 0.33 | 3.6 | 55692 | 0.3340 |
73
+ | 0.3283 | 3.8 | 58786 | 0.3282 |
74
+ | 0.3266 | 4.0 | 61880 | 0.3166 |
75
+ | 0.317 | 4.2 | 64974 | 0.3149 |
76
+ | 0.3122 | 4.4 | 68068 | 0.3149 |
77
+ | 0.313 | 4.6 | 71162 | 0.3147 |
78
+ | 0.3043 | 4.8 | 74256 | 0.3130 |
79
+ | 0.3019 | 5.0 | 77350 | 0.3036 |
80
+ | 0.2952 | 5.2 | 80444 | 0.3000 |
81
+ | 0.2996 | 5.4 | 83538 | 0.3003 |
82
+ | 0.2957 | 5.6 | 86632 | 0.2993 |
83
+ | 0.2935 | 5.8 | 89726 | 0.3047 |
84
+ | 0.2885 | 6.0 | 92820 | 0.2928 |
85
+ | 0.2755 | 6.2 | 95914 | 0.2915 |
86
+ | 0.2763 | 6.4 | 99008 | 0.2875 |
87
+ | 0.2755 | 6.6 | 102102 | 0.2855 |
88
+ | 0.2811 | 6.8 | 105196 | 0.2812 |
89
+ | 0.2704 | 7.0 | 108290 | 0.2796 |
90
+ | 0.26 | 7.2 | 111384 | 0.2776 |
91
+ | 0.2564 | 7.4 | 114478 | 0.2691 |
92
+ | 0.2613 | 7.6 | 117572 | 0.2702 |
93
+ | 0.2568 | 7.8 | 120666 | 0.2684 |
94
+ | 0.2579 | 8.0 | 123760 | 0.2643 |
95
+ | 0.2422 | 8.2 | 126854 | 0.2624 |
96
+ | 0.243 | 8.4 | 129948 | 0.2619 |
97
+ | 0.2421 | 8.6 | 133042 | 0.2583 |
98
+ | 0.2455 | 8.8 | 136136 | 0.2575 |
99
+ | 0.2428 | 9.0 | 139230 | 0.2511 |
100
+ | 0.2286 | 9.2 | 142324 | 0.2478 |
101
+ | 0.227 | 9.4 | 145418 | 0.2507 |
102
+ | 0.2246 | 9.6 | 148512 | 0.2474 |
103
+ | 0.2273 | 9.8 | 151606 | 0.2452 |
104
+ | 0.2211 | 10.0 | 154700 | 0.2432 |
105
+ | 0.2117 | 10.2 | 157794 | 0.2434 |
106
+ | 0.2098 | 10.4 | 160888 | 0.2377 |
107
+ | 0.2092 | 10.6 | 163982 | 0.2376 |
108
+ | 0.2073 | 10.8 | 167076 | 0.2355 |
109
+ | 0.2051 | 11.0 | 170170 | 0.2303 |
110
+ | 0.1966 | 11.2 | 173264 | 0.2321 |
111
+ | 0.1923 | 11.4 | 176358 | 0.2294 |
112
+ | 0.1913 | 11.6 | 179452 | 0.2275 |
113
+ | 0.189 | 11.8 | 182546 | 0.2267 |
114
+ | 0.1924 | 12.0 | 185640 | 0.2261 |
115
 
116
 
117
  ### Framework versions
adapter_config.json CHANGED
@@ -23,8 +23,8 @@
23
  "rank_pattern": {},
24
  "revision": null,
25
  "target_modules": [
26
- "v_proj",
27
- "q_proj"
28
  ],
29
  "task_type": "CAUSAL_LM",
30
  "use_dora": false,
 
23
  "rank_pattern": {},
24
  "revision": null,
25
  "target_modules": [
26
+ "q_proj",
27
+ "v_proj"
28
  ],
29
  "task_type": "CAUSAL_LM",
30
  "use_dora": false,
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:308667600ebb9fd4fe2b2cd4523192f26254ad5366e4c3753bf37c0124ecceae
3
  size 268636736
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:22253574595cd04088755a3c511c1b3722206ed2cdaf7dba620e1b61cdffcf5b
3
  size 268636736
loss_plot.png CHANGED
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5333ba3480d2d9eeb1f6cc9b990ec9d7d7599681ca30c253e6779c7cbb7010ba
3
- size 5496
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3476aecc2e7edfbf65feed98e3e459c9ff12870258371744a90a8e58d9761d2e
3
+ size 5432