Hugging Face's logo

colorlessideas
/
xls-r-300m-mocho-120

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
wav2vec2
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
colorlessideas commited on
Commit
111256d
·
verified ·
1 Parent(s): 71d269d

End of training

Browse files
Files changed (2) hide show
  1. README.md +13 -23
  2. model.safetensors +1 -1
README.md CHANGED
@@ -16,8 +16,8 @@ should probably proofread and complete it, then remove this comment. -->
16
 
17
  This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the None dataset.
18
  It achieves the following results on the evaluation set:
19
- - Loss: 0.0414
20
- - Cer: 0.0204
21
 
22
  ## Model description
23
 
@@ -36,12 +36,12 @@ More information needed
36
  ### Training hyperparameters
37
 
38
  The following hyperparameters were used during training:
39
- - learning_rate: 0.0001
40
- - train_batch_size: 8
41
  - eval_batch_size: 8
42
  - seed: 42
43
  - gradient_accumulation_steps: 2
44
- - total_train_batch_size: 16
45
  - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
46
  - lr_scheduler_type: linear
47
  - lr_scheduler_warmup_steps: 500
@@ -52,24 +52,14 @@ The following hyperparameters were used during training:
52
 
53
  | Training Loss | Epoch | Step | Validation Loss | Cer |
54
  |:-------------:|:-------:|:----:|:---------------:|:------:|
55
- | 6.3894 | 1.5822 | 400 | 2.9912 | 0.9993 |
56
- | 1.8056 | 3.1624 | 800 | 0.8697 | 0.2591 |
57
- | 1.057 | 4.7446 | 1200 | 0.6070 | 0.1843 |
58
- | 0.8359 | 6.3248 | 1600 | 0.4518 | 0.1386 |
59
- | 0.6848 | 7.9069 | 2000 | 0.3494 | 0.1075 |
60
- | 0.5899 | 9.4871 | 2400 | 0.2883 | 0.0898 |
61
- | 0.5033 | 11.0673 | 2800 | 0.2550 | 0.0768 |
62
- | 0.4508 | 12.6495 | 3200 | 0.2317 | 0.0714 |
63
- | 0.39 | 14.2297 | 3600 | 0.2030 | 0.0614 |
64
- | 0.3583 | 15.8119 | 4000 | 0.1736 | 0.0577 |
65
- | 0.3038 | 17.3921 | 4400 | 0.1573 | 0.0475 |
66
- | 0.2891 | 18.9743 | 4800 | 0.1310 | 0.0448 |
67
- | 0.2488 | 20.5545 | 5200 | 0.1233 | 0.0387 |
68
- | 0.2254 | 22.1347 | 5600 | 0.1062 | 0.0327 |
69
- | 0.1936 | 23.7168 | 6000 | 0.0811 | 0.0305 |
70
- | 0.1638 | 25.2970 | 6400 | 0.0641 | 0.0254 |
71
- | 0.1489 | 26.8792 | 6800 | 0.0499 | 0.0211 |
72
- | 0.1361 | 28.4594 | 7200 | 0.0414 | 0.0204 |
73
 
74
 
75
  ### Framework versions
 
16
 
17
  This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the None dataset.
18
  It achieves the following results on the evaluation set:
19
+ - Loss: 0.0505
20
+ - Cer: 0.0300
21
 
22
  ## Model description
23
 
 
36
  ### Training hyperparameters
37
 
38
  The following hyperparameters were used during training:
39
+ - learning_rate: 0.001
40
+ - train_batch_size: 16
41
  - eval_batch_size: 8
42
  - seed: 42
43
  - gradient_accumulation_steps: 2
44
+ - total_train_batch_size: 32
45
  - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
46
  - lr_scheduler_type: linear
47
  - lr_scheduler_warmup_steps: 500
 
52
 
53
  | Training Loss | Epoch | Step | Validation Loss | Cer |
54
  |:-------------:|:-------:|:----:|:---------------:|:------:|
55
+ | 3.3995 | 3.6364 | 400 | 1.1680 | 0.4202 |
56
+ | 1.3772 | 7.2727 | 800 | 0.7910 | 0.3031 |
57
+ | 1.0304 | 10.9091 | 1200 | 0.5138 | 0.1994 |
58
+ | 0.8008 | 14.5455 | 1600 | 0.3396 | 0.1475 |
59
+ | 0.6073 | 18.1818 | 2000 | 0.2182 | 0.0974 |
60
+ | 0.4772 | 21.8182 | 2400 | 0.1480 | 0.0688 |
61
+ | 0.3533 | 25.4545 | 2800 | 0.0960 | 0.0500 |
62
+ | 0.2672 | 29.0909 | 3200 | 0.0505 | 0.0300 |
 
 
 
 
 
 
 
 
 
 
63
 
64
 
65
  ### Framework versions
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:accbe62c306af7dea4b406bb9c2e5563bed6245deb8c0d04c5c119f13f62fbf5
3
  size 1261942780
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4c40c7a4a8265c339d538121dc0bf57bd255c5b84cd9524046477e8b81b4d37c
3
  size 1261942780