Hugging Face's logo

miosipof
/
asr_EN_medium_v1

PEFT
TensorBoard
Safetensors
English
Generated from Trainer
Eval Results
Model card Files Metrics Community
miosipof commited on
Commit
3a2783e
•
1 Parent(s): c326954

End of training

Browse files
Files changed (5) hide show
  1. README.md +39 -23
  2. adapter_model.safetensors +1 -1
  3. runs/Oct02_18-11-01_fd95e0b5707e/events.out.tfevents.1727892677.fd95e0b5707e +3 -0
  4. runs/Oct02_18-15-50_fd95e0b5707e/events.out.tfevents.1727892977.fd95e0b5707e +3 -0
  5. training_args.bin +1 -1
README.md CHANGED
@@ -24,7 +24,7 @@ model-index:
24
  args: default
25
  metrics:
26
  - type: wer
27
- value: 106.7911714770798
28
  name: Wer
29
  ---
30
 
@@ -35,8 +35,8 @@ should probably proofread and complete it, then remove this comment. -->
35
 
36
  This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the miosipof/asr_en dataset.
37
  It achieves the following results on the evaluation set:
38
- - Loss: 3.9473
39
- - Wer: 106.7912
40
 
41
  ## Model description
42
 
@@ -55,15 +55,15 @@ More information needed
55
  ### Training hyperparameters
56
 
57
  The following hyperparameters were used during training:
58
- - learning_rate: 1e-05
59
- - train_batch_size: 32
60
  - eval_batch_size: 16
61
  - seed: 42
62
  - gradient_accumulation_steps: 2
63
- - total_train_batch_size: 64
64
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
65
  - lr_scheduler_type: linear
66
- - lr_scheduler_warmup_steps: 128
67
  - training_steps: 1024
68
  - mixed_precision_training: Native AMP
69
 
@@ -71,22 +71,38 @@ The following hyperparameters were used during training:
71
 
72
  | Training Loss | Epoch | Step | Validation Loss | Wer |
73
  |:-------------:|:-------:|:----:|:---------------:|:--------:|
74
- | 7.4432 | 4.2667 | 64 | 7.9570 | 167.9117 |
75
- | 6.9871 | 8.5333 | 128 | 7.0858 | 167.5722 |
76
- | 6.1972 | 12.8 | 192 | 6.3333 | 205.2632 |
77
- | 5.9006 | 17.0667 | 256 | 6.0843 | 203.5654 |
78
- | 5.61 | 21.3333 | 320 | 5.8153 | 168.7606 |
79
- | 5.2344 | 25.6 | 384 | 5.4746 | 168.0815 |
80
- | 4.8067 | 29.8667 | 448 | 5.0913 | 168.2513 |
81
- | 4.3927 | 34.1333 | 512 | 4.7586 | 201.5280 |
82
- | 4.1103 | 38.4 | 576 | 4.5158 | 164.5161 |
83
- | 3.8975 | 42.6667 | 640 | 4.3460 | 108.4890 |
84
- | 3.7471 | 46.9333 | 704 | 4.2178 | 109.1681 |
85
- | 3.6146 | 51.2 | 768 | 4.1226 | 108.3192 |
86
- | 3.53 | 55.4667 | 832 | 4.0471 | 107.9796 |
87
- | 3.4579 | 59.7333 | 896 | 3.9927 | 107.4703 |
88
- | 3.4061 | 64.0 | 960 | 3.9594 | 106.9610 |
89
- | 3.3577 | 68.2667 | 1024 | 3.9473 | 106.7912 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
90
 
91
 
92
  ### Framework versions
 
24
  args: default
25
  metrics:
26
  - type: wer
27
+ value: 68.88888888888889
28
  name: Wer
29
  ---
30
 
 
35
 
36
  This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the miosipof/asr_en dataset.
37
  It achieves the following results on the evaluation set:
38
+ - Loss: 0.3857
39
+ - Wer: 68.8889
40
 
41
  ## Model description
42
 
 
55
  ### Training hyperparameters
56
 
57
  The following hyperparameters were used during training:
58
+ - learning_rate: 0.001
59
+ - train_batch_size: 16
60
  - eval_batch_size: 16
61
  - seed: 42
62
  - gradient_accumulation_steps: 2
63
+ - total_train_batch_size: 32
64
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
65
  - lr_scheduler_type: linear
66
+ - lr_scheduler_warmup_steps: 32
67
  - training_steps: 1024
68
  - mixed_precision_training: Native AMP
69
 
 
71
 
72
  | Training Loss | Epoch | Step | Validation Loss | Wer |
73
  |:-------------:|:-------:|:----:|:---------------:|:--------:|
74
+ | 5.6529 | 1.0847 | 32 | 2.0524 | 41.1111 |
75
+ | 1.2225 | 2.1695 | 64 | 0.6093 | 52.0635 |
76
+ | 0.2888 | 3.2542 | 96 | 0.4636 | 48.5714 |
77
+ | 0.1701 | 4.3390 | 128 | 0.4190 | 43.0159 |
78
+ | 0.1729 | 5.4237 | 160 | 0.5561 | 61.9048 |
79
+ | 0.0846 | 6.5085 | 192 | 0.3515 | 57.3016 |
80
+ | 0.0678 | 7.5932 | 224 | 0.3795 | 47.9365 |
81
+ | 0.0578 | 8.6780 | 256 | 0.5905 | 56.9841 |
82
+ | 0.0457 | 9.7627 | 288 | 0.4444 | 73.0159 |
83
+ | 0.0432 | 10.8475 | 320 | 0.5010 | 59.2063 |
84
+ | 0.0407 | 11.9322 | 352 | 0.5758 | 63.4921 |
85
+ | 0.0341 | 13.0169 | 384 | 0.6487 | 50.3175 |
86
+ | 0.0308 | 14.1017 | 416 | 0.4682 | 45.8730 |
87
+ | 0.0304 | 15.1864 | 448 | 0.4518 | 65.5556 |
88
+ | 0.0241 | 16.2712 | 480 | 0.5138 | 64.2857 |
89
+ | 0.029 | 17.3559 | 512 | 0.5460 | 66.5079 |
90
+ | 0.0169 | 18.4407 | 544 | 0.6139 | 64.7619 |
91
+ | 0.0196 | 19.5254 | 576 | 0.6055 | 54.4444 |
92
+ | 0.0148 | 20.6102 | 608 | 0.4502 | 65.7143 |
93
+ | 0.0153 | 21.6949 | 640 | 0.4179 | 81.7460 |
94
+ | 0.0149 | 22.7797 | 672 | 0.4491 | 108.7302 |
95
+ | 0.0188 | 23.8644 | 704 | 0.3885 | 75.3968 |
96
+ | 0.0115 | 24.9492 | 736 | 0.4070 | 182.6984 |
97
+ | 0.0111 | 26.0339 | 768 | 0.4429 | 128.7302 |
98
+ | 0.0124 | 27.1186 | 800 | 0.3827 | 69.2063 |
99
+ | 0.0096 | 28.2034 | 832 | 0.4028 | 70.0 |
100
+ | 0.0121 | 29.2881 | 864 | 0.3651 | 63.8095 |
101
+ | 0.0083 | 30.3729 | 896 | 0.3906 | 66.6667 |
102
+ | 0.0085 | 31.4576 | 928 | 0.3861 | 66.8254 |
103
+ | 0.0092 | 32.5424 | 960 | 0.3834 | 69.6825 |
104
+ | 0.0095 | 33.6271 | 992 | 0.3861 | 68.8889 |
105
+ | 0.007 | 34.7119 | 1024 | 0.3857 | 68.8889 |
106
 
107
 
108
  ### Framework versions
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1aeb882533eba03b542ed0bb46eb9db32b7dc782bd6fee502cc8a8fc6d583cf9
3
  size 18915424
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3d21652d7d1290ca5032878a69d8081441891ea1b33e81ea630ad9b7e51554c1
3
  size 18915424
runs/Oct02_18-11-01_fd95e0b5707e/events.out.tfevents.1727892677.fd95e0b5707e ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:015d3b05a98070451da5d1aa144c347292aa58ab815bc6aabb75541da331a853
3
+ size 4320
runs/Oct02_18-15-50_fd95e0b5707e/events.out.tfevents.1727892977.fd95e0b5707e ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5d8e16cdfdac92b8a40177f64bc34e8dbfbf773333e5a04d4c99bd8fdd3fb596
3
+ size 23080
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:23815f753814c6fa06c941b85aa50078574dca65e331b33048cc31379b12fea2
3
  size 5304
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5d8e304d4a0ab457402ff42f37ff7b96f2986612b0a5cbf51241b59b06eca3bc
3
  size 5304