Hugging Face's logo

aydink
/
unispeech-sat-base-finetuned-common_voice

Audio Classification
Transformers
TensorBoard
Safetensors
unispeech-sat
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
aydink commited on
Commit
da6d21d
•
1 Parent(s): 8732e91

Model save

Browse files
Files changed (2) hide show
  1. README.md +20 -22
  2. model.safetensors +1 -1
README.md CHANGED
@@ -19,13 +19,13 @@ should probably proofread and complete it, then remove this comment. -->
19
 
20
  This model is a fine-tuned version of [microsoft/unispeech-sat-base](https://huggingface.co/microsoft/unispeech-sat-base) on the None dataset.
21
  It achieves the following results on the evaluation set:
22
- - Loss: 0.0013
23
- - Accuracy: 1.0
24
- - F1: 1.0
25
- - Recall: 1.0
26
- - Precision: 1.0
27
- - Mcc: 1.0
28
- - Auc: 1.0
29
 
30
  ## Model description
31
 
@@ -44,12 +44,10 @@ More information needed
44
  ### Training hyperparameters
45
 
46
  The following hyperparameters were used during training:
47
- - learning_rate: 3e-05
48
  - train_batch_size: 8
49
  - eval_batch_size: 8
50
  - seed: 42
51
- - gradient_accumulation_steps: 4
52
- - total_train_batch_size: 32
53
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
54
  - lr_scheduler_type: linear
55
  - lr_scheduler_warmup_ratio: 0.1
@@ -60,21 +58,21 @@ The following hyperparameters were used during training:
60
 
61
  | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Recall | Precision | Mcc | Auc |
62
  |:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|:------:|:---------:|:------:|:------:|
63
- | 0.1326 | 1.0 | 50 | 0.0283 | 0.9975 | 0.9975 | 0.9975 | 0.9975 | 0.9969 | 1.0 |
64
- | 0.1323 | 2.0 | 100 | 0.0450 | 0.9925 | 0.9925 | 0.9925 | 0.9928 | 0.9907 | 0.9994 |
65
- | 0.0499 | 3.0 | 150 | 0.0081 | 0.9975 | 0.9975 | 0.9975 | 0.9975 | 0.9969 | 1.0 |
66
- | 0.0183 | 4.0 | 200 | 0.0037 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 |
67
- | 0.024 | 5.0 | 250 | 0.0026 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 |
68
- | 0.0326 | 6.0 | 300 | 0.0020 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 |
69
- | 0.0172 | 7.0 | 350 | 0.0017 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 |
70
- | 0.0053 | 8.0 | 400 | 0.0015 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 |
71
- | 0.0034 | 9.0 | 450 | 0.0050 | 0.9975 | 0.9975 | 0.9975 | 0.9975 | 0.9969 | 1.0 |
72
- | 0.014 | 10.0 | 500 | 0.0013 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 |
73
 
74
 
75
  ### Framework versions
76
 
77
- - Transformers 4.40.2
78
- - Pytorch 2.2.1+cu121
79
  - Datasets 2.19.1
80
  - Tokenizers 0.19.1
 
19
 
20
  This model is a fine-tuned version of [microsoft/unispeech-sat-base](https://huggingface.co/microsoft/unispeech-sat-base) on the None dataset.
21
  It achieves the following results on the evaluation set:
22
+ - Loss: 0.1896
23
+ - Accuracy: 0.96
24
+ - F1: 0.9601
25
+ - Recall: 0.96
26
+ - Precision: 0.9606
27
+ - Mcc: 0.9501
28
+ - Auc: 0.9939
29
 
30
  ## Model description
31
 
 
44
  ### Training hyperparameters
45
 
46
  The following hyperparameters were used during training:
47
+ - learning_rate: 1e-05
48
  - train_batch_size: 8
49
  - eval_batch_size: 8
50
  - seed: 42
 
 
51
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
52
  - lr_scheduler_type: linear
53
  - lr_scheduler_warmup_ratio: 0.1
 
58
 
59
  | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Recall | Precision | Mcc | Auc |
60
  |:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|:------:|:---------:|:------:|:------:|
61
+ | 1.5599 | 1.0 | 200 | 1.5446 | 0.415 | 0.3951 | 0.4150 | 0.6762 | 0.3213 | 0.8445 |
62
+ | 1.1707 | 2.0 | 400 | 1.0171 | 0.7575 | 0.7502 | 0.7575 | 0.7665 | 0.7023 | 0.9487 |
63
+ | 0.7857 | 3.0 | 600 | 0.7125 | 0.8375 | 0.8311 | 0.8375 | 0.8453 | 0.8008 | 0.9667 |
64
+ | 0.5713 | 4.0 | 800 | 0.5097 | 0.88 | 0.8794 | 0.8800 | 0.8929 | 0.8536 | 0.9874 |
65
+ | 0.4225 | 5.0 | 1000 | 0.3919 | 0.9075 | 0.9076 | 0.9075 | 0.9116 | 0.8853 | 0.9894 |
66
+ | 0.5846 | 6.0 | 1200 | 0.3119 | 0.9325 | 0.9327 | 0.9325 | 0.9355 | 0.9163 | 0.9883 |
67
+ | 0.3004 | 7.0 | 1400 | 0.2308 | 0.9475 | 0.9477 | 0.9475 | 0.9487 | 0.9346 | 0.9925 |
68
+ | 0.3011 | 8.0 | 1600 | 0.1974 | 0.955 | 0.9551 | 0.9550 | 0.9557 | 0.9439 | 0.9940 |
69
+ | 0.138 | 9.0 | 1800 | 0.1851 | 0.96 | 0.9601 | 0.96 | 0.9606 | 0.9501 | 0.9932 |
70
+ | 0.1582 | 10.0 | 2000 | 0.1896 | 0.96 | 0.9601 | 0.96 | 0.9606 | 0.9501 | 0.9939 |
71
 
72
 
73
  ### Framework versions
74
 
75
+ - Transformers 4.41.0
76
+ - Pytorch 2.3.0+cu121
77
  - Datasets 2.19.1
78
  - Tokenizers 0.19.1
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:10e9e6f22b3c8406a13e9562d9638a5ac3ac668de59413530e51144c2c5d1ed3
3
  size 378306500
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:63a8063fc701f0d403d6fd066c878ff4bb8110293ae57d1973ecf9c67339efb0
3
  size 378306500