Hugging Face's logo

yuriyvnv
/
whisper-tiny-mixed-pt

Automatic Speech Recognition
Transformers
Safetensors
Portuguese
whisper
portuguese
speech
audio
synthetic-data
asr
hf-asr-leaderboard
Eval Results (legacy)
Model card Files
xet
 Community
yuriyvnv commited on
Commit
3b1228e
·
verified ·
1 Parent(s): a2cc228

Model save

Browse files
Files changed (3) hide show
  1. README.md +36 -20
  2. final_model/model.safetensors +1 -1
  3. final_model/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 [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) on an unknown dataset.
18
  It achieves the following results on the evaluation set:
19
- - Loss: 0.6277
20
 
21
  ## Model description
22
 
@@ -35,35 +35,51 @@ More information needed
35
  ### Training hyperparameters
36
 
37
  The following hyperparameters were used during training:
38
- - learning_rate: 1e-06
39
  - train_batch_size: 256
40
  - eval_batch_size: 8
41
  - seed: 42
42
  - optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
43
  - lr_scheduler_type: linear
44
- - lr_scheduler_warmup_steps: 80
45
- - training_steps: 810
46
 
47
  ### Training results
48
 
49
  | Training Loss | Epoch | Step | Validation Loss |
50
  |:-------------:|:------:|:----:|:---------------:|
51
- | 1.4147 | 0.3106 | 50 | 1.2914 |
52
- | 1.1229 | 0.6211 | 100 | 0.9808 |
53
- | 0.7911 | 0.9317 | 150 | 0.7261 |
54
- | 0.7089 | 1.2422 | 200 | 0.6867 |
55
- | 0.6718 | 1.5528 | 250 | 0.6710 |
56
- | 0.6777 | 1.8634 | 300 | 0.6601 |
57
- | 0.6413 | 2.1739 | 350 | 0.6524 |
58
- | 0.626 | 2.4845 | 400 | 0.6468 |
59
- | 0.6065 | 2.7950 | 450 | 0.6416 |
60
- | 0.6059 | 3.1056 | 500 | 0.6373 |
61
- | 0.5958 | 3.4161 | 550 | 0.6344 |
62
- | 0.6057 | 3.7267 | 600 | 0.6318 |
63
- | 0.5941 | 4.0373 | 650 | 0.6301 |
64
- | 0.5965 | 4.3478 | 700 | 0.6290 |
65
- | 0.5871 | 4.6584 | 750 | 0.6281 |
66
- | 0.5993 | 4.9689 | 800 | 0.6277 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
67
 
68
 
69
  ### Framework versions
 
16
 
17
  This model is a fine-tuned version of [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) on an unknown dataset.
18
  It achieves the following results on the evaluation set:
19
+ - Loss: 0.5099
20
 
21
  ## Model description
22
 
 
35
  ### Training hyperparameters
36
 
37
  The following hyperparameters were used during training:
38
+ - learning_rate: 5e-06
39
  - train_batch_size: 256
40
  - eval_batch_size: 8
41
  - seed: 42
42
  - optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
43
  - lr_scheduler_type: linear
44
+ - lr_scheduler_warmup_ratio: 0.1
45
+ - num_epochs: 10
46
 
47
  ### Training results
48
 
49
  | Training Loss | Epoch | Step | Validation Loss |
50
  |:-------------:|:------:|:----:|:---------------:|
51
+ | 1.3211 | 0.3106 | 50 | 1.1319 |
52
+ | 0.7848 | 0.6211 | 100 | 0.6933 |
53
+ | 0.6263 | 0.9317 | 150 | 0.6365 |
54
+ | 0.5572 | 1.2422 | 200 | 0.6066 |
55
+ | 0.5236 | 1.5528 | 250 | 0.5880 |
56
+ | 0.5247 | 1.8634 | 300 | 0.5733 |
57
+ | 0.4671 | 2.1739 | 350 | 0.5649 |
58
+ | 0.4572 | 2.4845 | 400 | 0.5555 |
59
+ | 0.4383 | 2.7950 | 450 | 0.5487 |
60
+ | 0.4204 | 3.1056 | 500 | 0.5429 |
61
+ | 0.4085 | 3.4161 | 550 | 0.5390 |
62
+ | 0.4136 | 3.7267 | 600 | 0.5340 |
63
+ | 0.3966 | 4.0373 | 650 | 0.5297 |
64
+ | 0.3827 | 4.3478 | 700 | 0.5280 |
65
+ | 0.3799 | 4.6584 | 750 | 0.5259 |
66
+ | 0.3861 | 4.9689 | 800 | 0.5229 |
67
+ | 0.356 | 5.2795 | 850 | 0.5212 |
68
+ | 0.3608 | 5.5901 | 900 | 0.5191 |
69
+ | 0.3555 | 5.9006 | 950 | 0.5175 |
70
+ | 0.3436 | 6.2112 | 1000 | 0.5167 |
71
+ | 0.3437 | 6.5217 | 1050 | 0.5153 |
72
+ | 0.3428 | 6.8323 | 1100 | 0.5140 |
73
+ | 0.3352 | 7.1429 | 1150 | 0.5126 |
74
+ | 0.3343 | 7.4534 | 1200 | 0.5125 |
75
+ | 0.3281 | 7.7640 | 1250 | 0.5118 |
76
+ | 0.318 | 8.0745 | 1300 | 0.5115 |
77
+ | 0.329 | 8.3851 | 1350 | 0.5110 |
78
+ | 0.3271 | 8.6957 | 1400 | 0.5104 |
79
+ | 0.3215 | 9.0062 | 1450 | 0.5104 |
80
+ | 0.3318 | 9.3168 | 1500 | 0.5100 |
81
+ | 0.3175 | 9.6273 | 1550 | 0.5100 |
82
+ | 0.3148 | 9.9379 | 1600 | 0.5099 |
83
 
84
 
85
  ### Framework versions
final_model/model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:159d6a6b292d232ec4d0b595812f9475b267eb3baafa4d6b5c151c5b59427cbb
3
  size 151061672
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f231df0b854765ea9822cfde8a8f74044a91c9a2e36f725298baae7c2ebfa00d
3
  size 151061672
final_model/training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:f23e84d505250fe8fcb0603231d52908696c0de4c2eac2fcfc8ff0814a4c70ba
3
- size 5688
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8182887d6e1a6ec58abf5a883022b2904ab2e7463896b7dd56e19a3862ac7245
3
+ size 5624