Hugging Face's logo

xuliu15
/
English_32r_LoRA_1h

PEFT
TensorBoard
Safetensors
Generated from Trainer
Model card Files Metrics Community
xuliu15 commited on
Commit
c8712f6
•
1 Parent(s): 1cbc553

End of training

Browse files
Files changed (5) hide show
  1. README.md +56 -28
  2. adapter_config.json +3 -3
  3. adapter_model.safetensors +2 -2
  4. runs/May26_16-18-56_a100gpu5/events.out.tfevents.1716733156.a100gpu5.1523333.0 +3 -0
  5. training_args.bin +1 -1
README.md CHANGED
@@ -18,7 +18,7 @@ should probably proofread and complete it, then remove this comment. -->
18
 
19
  This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the librispeech dataset.
20
  It achieves the following results on the evaluation set:
21
- - Loss: 1.9444
22
 
23
  ## Model description
24
 
@@ -37,42 +37,70 @@ More information needed
37
  ### Training hyperparameters
38
 
39
  The following hyperparameters were used during training:
40
- - learning_rate: 1e-06
41
  - train_batch_size: 8
42
- - eval_batch_size: 4
43
  - seed: 42
44
- - gradient_accumulation_steps: 2
45
- - total_train_batch_size: 16
46
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
47
  - lr_scheduler_type: linear
48
  - lr_scheduler_warmup_steps: 50
49
- - num_epochs: 20
50
  - mixed_precision_training: Native AMP
51
 
52
  ### Training results
53
 
54
- | Training Loss | Epoch | Step | Validation Loss |
55
- |:-------------:|:-------:|:----:|:---------------:|
56
- | No log | 0.9744 | 19 | 2.1407 |
57
- | 1.3439 | 2.0 | 39 | 2.1284 |
58
- | 1.3611 | 2.9744 | 58 | 2.1101 |
59
- | 1.3218 | 4.0 | 78 | 2.0900 |
60
- | 1.3218 | 4.9744 | 97 | 2.0717 |
61
- | 1.3312 | 6.0 | 117 | 2.0546 |
62
- | 1.3108 | 6.9744 | 136 | 2.0389 |
63
- | 1.2979 | 8.0 | 156 | 2.0235 |
64
- | 1.2662 | 8.9744 | 175 | 2.0105 |
65
- | 1.2662 | 10.0 | 195 | 1.9979 |
66
- | 1.2706 | 10.9744 | 214 | 1.9876 |
67
- | 1.2673 | 12.0 | 234 | 1.9781 |
68
- | 1.2302 | 12.9744 | 253 | 1.9695 |
69
- | 1.2302 | 14.0 | 273 | 1.9620 |
70
- | 1.2478 | 14.9744 | 292 | 1.9564 |
71
- | 1.2536 | 16.0 | 312 | 1.9516 |
72
- | 1.2125 | 16.9744 | 331 | 1.9480 |
73
- | 1.2406 | 18.0 | 351 | 1.9457 |
74
- | 1.2406 | 18.9744 | 370 | 1.9445 |
75
- | 1.218 | 19.4872 | 380 | 1.9444 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
76
 
77
 
78
  ### Framework versions
 
18
 
19
  This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the librispeech dataset.
20
  It achieves the following results on the evaluation set:
21
+ - Loss: 0.3666
22
 
23
  ## Model description
24
 
 
37
  ### Training hyperparameters
38
 
39
  The following hyperparameters were used during training:
40
+ - learning_rate: 1e-05
41
  - train_batch_size: 8
42
+ - eval_batch_size: 8
43
  - seed: 42
 
 
44
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
45
  - lr_scheduler_type: linear
46
  - lr_scheduler_warmup_steps: 50
47
+ - num_epochs: 50
48
  - mixed_precision_training: Native AMP
49
 
50
  ### Training results
51
 
52
+ | Training Loss | Epoch | Step | Validation Loss |
53
+ |:-------------:|:-----:|:----:|:---------------:|
54
+ | 1.8938 | 1.0 | 39 | 2.5946 |
55
+ | 1.4637 | 2.0 | 78 | 2.0459 |
56
+ | 1.3003 | 3.0 | 117 | 1.6739 |
57
+ | 0.9415 | 4.0 | 156 | 1.2729 |
58
+ | 0.8165 | 5.0 | 195 | 1.0158 |
59
+ | 0.6326 | 6.0 | 234 | 0.9033 |
60
+ | 0.5716 | 7.0 | 273 | 0.7272 |
61
+ | 0.4662 | 8.0 | 312 | 0.6731 |
62
+ | 0.4133 | 9.0 | 351 | 0.6433 |
63
+ | 0.4 | 10.0 | 390 | 0.6248 |
64
+ | 0.3862 | 11.0 | 429 | 0.6103 |
65
+ | 0.3901 | 12.0 | 468 | 0.5962 |
66
+ | 0.3661 | 13.0 | 507 | 0.5841 |
67
+ | 0.3609 | 14.0 | 546 | 0.5739 |
68
+ | 0.3439 | 15.0 | 585 | 0.5660 |
69
+ | 0.3391 | 16.0 | 624 | 0.5581 |
70
+ | 0.3231 | 17.0 | 663 | 0.5510 |
71
+ | 0.3097 | 18.0 | 702 | 0.5441 |
72
+ | 0.2994 | 19.0 | 741 | 0.5373 |
73
+ | 0.2991 | 20.0 | 780 | 0.5304 |
74
+ | 0.2972 | 21.0 | 819 | 0.5240 |
75
+ | 0.2898 | 22.0 | 858 | 0.5187 |
76
+ | 0.2809 | 23.0 | 897 | 0.5142 |
77
+ | 0.2845 | 24.0 | 936 | 0.5119 |
78
+ | 0.269 | 25.0 | 975 | 0.5074 |
79
+ | 0.2721 | 26.0 | 1014 | 0.5033 |
80
+ | 0.2633 | 27.0 | 1053 | 0.5006 |
81
+ | 0.2623 | 28.0 | 1092 | 0.4984 |
82
+ | 0.2492 | 29.0 | 1131 | 0.4931 |
83
+ | 0.25 | 30.0 | 1170 | 0.4861 |
84
+ | 0.2479 | 31.0 | 1209 | 0.4833 |
85
+ | 0.2416 | 32.0 | 1248 | 0.4777 |
86
+ | 0.2356 | 33.0 | 1287 | 0.4794 |
87
+ | 0.2281 | 34.0 | 1326 | 0.4663 |
88
+ | 0.2191 | 35.0 | 1365 | 0.4605 |
89
+ | 0.2218 | 36.0 | 1404 | 0.4600 |
90
+ | 0.2078 | 37.0 | 1443 | 0.4545 |
91
+ | 0.2122 | 38.0 | 1482 | 0.4470 |
92
+ | 0.2076 | 39.0 | 1521 | 0.4510 |
93
+ | 0.2004 | 40.0 | 1560 | 0.4326 |
94
+ | 0.2004 | 41.0 | 1599 | 0.4280 |
95
+ | 0.1901 | 42.0 | 1638 | 0.4342 |
96
+ | 0.1856 | 43.0 | 1677 | 0.4107 |
97
+ | 0.1802 | 44.0 | 1716 | 0.4060 |
98
+ | 0.1677 | 45.0 | 1755 | 0.4029 |
99
+ | 0.1658 | 46.0 | 1794 | 0.3922 |
100
+ | 0.1589 | 47.0 | 1833 | 0.3845 |
101
+ | 0.152 | 48.0 | 1872 | 0.3790 |
102
+ | 0.1493 | 49.0 | 1911 | 0.3691 |
103
+ | 0.1426 | 50.0 | 1950 | 0.3666 |
104
 
105
 
106
  ### Framework versions
adapter_config.json CHANGED
@@ -1,10 +1,10 @@
1
  {
2
  "alpha_pattern": {},
3
  "auto_mapping": {
4
- "base_model_class": "PeftModel",
5
- "parent_library": "peft.peft_model"
6
  },
7
- "base_model_name_or_path": null,
8
  "bias": "none",
9
  "fan_in_fan_out": false,
10
  "inference_mode": true,
 
1
  {
2
  "alpha_pattern": {},
3
  "auto_mapping": {
4
+ "base_model_class": "WhisperForConditionalGeneration",
5
+ "parent_library": "transformers.models.whisper.modeling_whisper"
6
  },
7
+ "base_model_name_or_path": "openai/whisper-small",
8
  "bias": "none",
9
  "fan_in_fan_out": false,
10
  "inference_mode": true,
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:486cad9f40dfd9aefd9d6b939e358a13f254767337ccb19a25fd97c22b9da3ff
3
- size 14180960
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b9d49b4232441c045eace3f9bbae8ee26b418f25fbd4c38f68ebb5c975d41e4f
3
+ size 14176064
runs/May26_16-18-56_a100gpu5/events.out.tfevents.1716733156.a100gpu5.1523333.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d978f8fdab9dc2a5e467ed2afbaac07afebc286bc4a04861417b2a6baf79e54a
3
+ size 35847
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:89f085cd7c7e547bba05ebd36d8e7c5fb02a5fb625e3568a4f5bf76a33c07549
3
  size 5176
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:91d76de232c9d1e19f7639da4d787ed3167d12180e0131921a845e6db4b131ec
3
  size 5176