Hugging Face's logo

MelanieKoe
/
w2v2-base-pretrained_lr5e-5_at1_da1

Automatic Speech Recognition
Transformers
Safetensors
wav2vec2
Generated from Trainer
Inference Endpoints
Model card Files Community
MelanieKoe commited on
Commit
87dba5c
•
1 Parent(s): bce40da

Model save

Browse files
Files changed (2) hide show
  1. README.md +19 -19
  2. model.safetensors +1 -1
README.md CHANGED
@@ -17,8 +17,8 @@ should probably proofread and complete it, then remove this comment. -->
17
 
18
  This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the None dataset.
19
  It achieves the following results on the evaluation set:
20
- - Loss: 2.2991
21
- - Wer: 0.3016
22
 
23
  ## Model description
24
 
@@ -43,7 +43,7 @@ The following hyperparameters were used during training:
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: 500
47
  - training_steps: 4000
48
  - mixed_precision_training: Native AMP
49
 
@@ -51,22 +51,22 @@ The following hyperparameters were used during training:
51
 
52
  | Training Loss | Epoch | Step | Validation Loss | Wer |
53
  |:-------------:|:-----:|:----:|:---------------:|:------:|
54
- | 17.3296 | 6.1 | 250 | 8.4283 | 1.0 |
55
- | 4.6394 | 12.2 | 500 | 7.5283 | 1.0 |
56
- | 3.4235 | 18.29 | 750 | 5.8716 | 1.0 |
57
- | 3.3011 | 24.39 | 1000 | 4.9842 | 1.0 |
58
- | 3.048 | 30.49 | 1250 | 3.1389 | 1.0077 |
59
- | 2.1291 | 36.59 | 1500 | 2.0209 | 0.9257 |
60
- | 0.5553 | 42.68 | 1750 | 1.5785 | 0.4742 |
61
- | 0.2117 | 48.78 | 2000 | 1.7978 | 0.3631 |
62
- | 0.1228 | 54.88 | 2250 | 1.8723 | 0.3323 |
63
- | 0.0771 | 60.98 | 2500 | 2.0055 | 0.3234 |
64
- | 0.0587 | 67.07 | 2750 | 2.0493 | 0.3157 |
65
- | 0.0398 | 73.17 | 3000 | 2.0708 | 0.3084 |
66
- | 0.0336 | 79.27 | 3250 | 2.2532 | 0.3148 |
67
- | 0.0271 | 85.37 | 3500 | 2.2694 | 0.3016 |
68
- | 0.0234 | 91.46 | 3750 | 2.3538 | 0.3050 |
69
- | 0.0213 | 97.56 | 4000 | 2.2991 | 0.3016 |
70
 
71
 
72
  ### Framework versions
 
17
 
18
  This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the None dataset.
19
  It achieves the following results on the evaluation set:
20
+ - Loss: 2.1337
21
+ - Wer: 0.1867
22
 
23
  ## Model description
24
 
 
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: 1000
47
  - training_steps: 4000
48
  - mixed_precision_training: Native AMP
49
 
 
51
 
52
  | Training Loss | Epoch | Step | Validation Loss | Wer |
53
  |:-------------:|:-----:|:----:|:---------------:|:------:|
54
+ | 24.34 | 6.1 | 250 | 4.9035 | 1.0 |
55
+ | 3.488 | 12.2 | 500 | 3.2532 | 1.0 |
56
+ | 3.1188 | 18.29 | 750 | 3.1635 | 1.0 |
57
+ | 3.074 | 24.39 | 1000 | 3.1097 | 1.0 |
58
+ | 3.0433 | 30.49 | 1250 | 3.0810 | 1.0 |
59
+ | 3.0198 | 36.59 | 1500 | 3.0607 | 1.0 |
60
+ | 1.554 | 42.68 | 1750 | 1.2500 | 0.3635 |
61
+ | 0.224 | 48.78 | 2000 | 1.6458 | 0.1999 |
62
+ | 0.1319 | 54.88 | 2250 | 1.6232 | 0.1982 |
63
+ | 0.0961 | 60.98 | 2500 | 1.8003 | 0.1845 |
64
+ | 0.0788 | 67.07 | 2750 | 2.0011 | 0.1939 |
65
+ | 0.0634 | 73.17 | 3000 | 2.0543 | 0.1909 |
66
+ | 0.0567 | 79.27 | 3250 | 2.0814 | 0.1867 |
67
+ | 0.0501 | 85.37 | 3500 | 2.0521 | 0.1828 |
68
+ | 0.0463 | 91.46 | 3750 | 2.1218 | 0.1837 |
69
+ | 0.0428 | 97.56 | 4000 | 2.1337 | 0.1867 |
70
 
71
 
72
  ### Framework versions
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0129d2dcf85c7bcb2798bdb460dacf650160b52d9d4d61a0d5c3dc8d0ac2f1bf
3
  size 377629528
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ce85e5693ccbedbcd92692f0242cbc6f0744e807fe6f1e9ada0e4bc4faacb878
3
  size 377629528