Hugging Face's logo

sfedar
/
distilhubert-finetuned-gtzan

Audio Classification
TensorBoard
Safetensors
hubert
Generated from Trainer
Eval Results
Model card Files Metrics Community
sfedar commited on
Commit
d16bdf4
•
1 Parent(s): 281a218

End of training, 20 epochs, 4 batch size, writer batch size: 500, 1 gradient accumulation steps, learning rate: 5e-05, 30 s

Browse files
Files changed (2) hide show
  1. README.md +31 -24
  2. model.safetensors +1 -1
README.md CHANGED
@@ -1,5 +1,4 @@
1
  ---
2
- library_name: transformers
3
  license: apache-2.0
4
  base_model: ntu-spml/distilhubert
5
  tags:
@@ -23,7 +22,7 @@ model-index:
23
  metrics:
24
  - name: Accuracy
25
  type: accuracy
26
- value: 0.75
27
  ---
28
 
29
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -33,8 +32,8 @@ should probably proofread and complete it, then remove this comment. -->
33
 
34
  This model is a fine-tuned version of [ntu-spml/distilhubert](https://huggingface.co/ntu-spml/distilhubert) on the GTZAN dataset.
35
  It achieves the following results on the evaluation set:
36
- - Loss: 0.8728
37
- - Accuracy: 0.75
38
 
39
  ## Model description
40
 
@@ -53,37 +52,45 @@ More information needed
53
  ### Training hyperparameters
54
 
55
  The following hyperparameters were used during training:
56
- - learning_rate: 9e-05
57
- - train_batch_size: 4
58
- - eval_batch_size: 4
59
  - seed: 42
60
- - gradient_accumulation_steps: 16
61
- - total_train_batch_size: 64
62
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
63
  - lr_scheduler_type: linear
64
  - lr_scheduler_warmup_ratio: 0.1
65
- - num_epochs: 10
66
  - mixed_precision_training: Native AMP
67
 
68
  ### Training results
69
 
70
- | Training Loss | Epoch | Step | Validation Loss | Accuracy |
71
- |:-------------:|:------:|:----:|:---------------:|:--------:|
72
- | 2.2872 | 0.9956 | 14 | 2.1804 | 0.31 |
73
- | 2.0327 | 1.9911 | 28 | 1.8048 | 0.55 |
74
- | 1.6638 | 2.9867 | 42 | 1.5831 | 0.57 |
75
- | 1.4257 | 3.9822 | 56 | 1.3604 | 0.66 |
76
- | 1.1899 | 4.9778 | 70 | 1.1437 | 0.71 |
77
- | 1.1012 | 5.9733 | 84 | 1.0711 | 0.69 |
78
- | 0.9901 | 6.9689 | 98 | 0.9877 | 0.72 |
79
- | 0.9088 | 7.9644 | 112 | 0.8899 | 0.79 |
80
- | 0.7938 | 8.96 | 126 | 0.8887 | 0.75 |
81
- | 0.8121 | 9.9556 | 140 | 0.8728 | 0.75 |
 
 
 
 
 
 
 
 
 
 
82
 
83
 
84
  ### Framework versions
85
 
86
- - Transformers 4.44.2
87
- - Pytorch 2.4.0+cu121
88
  - Datasets 2.21.0
89
  - Tokenizers 0.19.1
 
1
  ---
 
2
  license: apache-2.0
3
  base_model: ntu-spml/distilhubert
4
  tags:
 
22
  metrics:
23
  - name: Accuracy
24
  type: accuracy
25
+ value: 0.85
26
  ---
27
 
28
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
32
 
33
  This model is a fine-tuned version of [ntu-spml/distilhubert](https://huggingface.co/ntu-spml/distilhubert) on the GTZAN dataset.
34
  It achieves the following results on the evaluation set:
35
+ - Loss: 2.0298
36
+ - Accuracy: 0.85
37
 
38
  ## Model description
39
 
 
52
  ### Training hyperparameters
53
 
54
  The following hyperparameters were used during training:
55
+ - learning_rate: 5e-05
56
+ - train_batch_size: 8
57
+ - eval_batch_size: 8
58
  - seed: 42
 
 
59
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
60
  - lr_scheduler_type: linear
61
  - lr_scheduler_warmup_ratio: 0.1
62
+ - num_epochs: 20
63
  - mixed_precision_training: Native AMP
64
 
65
  ### Training results
66
 
67
+ | Training Loss | Epoch | Step | Validation Loss | Accuracy |
68
+ |:-------------:|:-----:|:----:|:---------------:|:--------:|
69
+ | 2.1346 | 1.0 | 113 | 2.0688 | 0.43 |
70
+ | 1.4627 | 2.0 | 226 | 1.5207 | 0.54 |
71
+ | 1.1848 | 3.0 | 339 | 1.2524 | 0.61 |
72
+ | 0.7007 | 4.0 | 452 | 0.9106 | 0.7 |
73
+ | 0.5724 | 5.0 | 565 | 0.7507 | 0.81 |
74
+ | 0.3465 | 6.0 | 678 | 0.6838 | 0.83 |
75
+ | 0.2814 | 7.0 | 791 | 0.5810 | 0.84 |
76
+ | 0.0838 | 8.0 | 904 | 0.7549 | 0.79 |
77
+ | 0.0866 | 9.0 | 1017 | 0.6639 | 0.82 |
78
+ | 0.0118 | 10.0 | 1130 | 0.8318 | 0.83 |
79
+ | 0.0019 | 11.0 | 1243 | 0.8335 | 0.84 |
80
+ | 0.0004 | 12.0 | 1356 | 1.2910 | 0.83 |
81
+ | 0.0 | 13.0 | 1469 | 1.3991 | 0.85 |
82
+ | 0.0 | 14.0 | 1582 | 1.7816 | 0.8 |
83
+ | 0.0 | 15.0 | 1695 | 1.8906 | 0.82 |
84
+ | 0.0 | 16.0 | 1808 | 2.0635 | 0.83 |
85
+ | 0.0 | 17.0 | 1921 | 1.9376 | 0.85 |
86
+ | 0.0 | 18.0 | 2034 | 2.0849 | 0.83 |
87
+ | 0.0 | 19.0 | 2147 | 2.0363 | 0.85 |
88
+ | 0.0 | 20.0 | 2260 | 2.0298 | 0.85 |
89
 
90
 
91
  ### Framework versions
92
 
93
+ - Transformers 4.44.0
94
+ - Pytorch 2.4.0
95
  - Datasets 2.21.0
96
  - Tokenizers 0.19.1
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:8284fe16522ad91d579c2352854f87407de150f5c0ba89e0e907cab32b987345
3
  size 94771728
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:95326786d63dc5fd97fafbb24e7c1f7c85e9a14494d87ce26a3e7579f4663342
3
  size 94771728