Hugging Face's logo

JamesJenkins
/
ast-finetuned-audioset-10-10-0.4593-finetuned-gtzan

Audio Classification
TensorBoard
Safetensors
audio-spectrogram-transformer
Generated from Trainer
Eval Results
Model card Files Metrics Community
JamesJenkins commited on
Commit
7e4dcbb
·
verified ·
1 Parent(s): e2d9a45

End of training

Browse files
Files changed (4) hide show
  1. README.md +21 -31
  2. model.safetensors +1 -1
  3. runs/Aug27_09-10-11_75209667f075/events.out.tfevents.1724749825.75209667f075.596.0 +3 -0
  4. training_args.bin +1 -1
README.md CHANGED
@@ -23,10 +23,10 @@ model-index:
23
  metrics:
24
  - name: Accuracy
25
  type: accuracy
26
- value: 0.86
27
  - name: F1
28
  type: f1
29
- value: 0.8599999999999999
30
  ---
31
 
32
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -36,9 +36,9 @@ should probably proofread and complete it, then remove this comment. -->
36
 
37
  This model is a fine-tuned version of [MIT/ast-finetuned-audioset-10-10-0.4593](https://huggingface.co/MIT/ast-finetuned-audioset-10-10-0.4593) on the GTZAN dataset.
38
  It achieves the following results on the evaluation set:
39
- - Loss: 0.7149
40
- - Accuracy: 0.86
41
- - F1: 0.8600
42
 
43
  ## Model description
44
 
@@ -61,43 +61,33 @@ The following hyperparameters were used during training:
61
  - train_batch_size: 2
62
  - eval_batch_size: 2
63
  - seed: 2024
64
- - gradient_accumulation_steps: 4
65
- - total_train_batch_size: 8
66
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
67
  - lr_scheduler_type: linear
68
  - lr_scheduler_warmup_ratio: 0.1
69
- - num_epochs: 20
70
  - mixed_precision_training: Native AMP
71
 
72
  ### Training results
73
 
74
- | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
75
- |:-------------:|:-------:|:----:|:---------------:|:--------:|:------:|
76
- | 0.6891 | 0.9956 | 112 | 0.6422 | 0.76 | 0.76 |
77
- | 0.7267 | 2.0 | 225 | 0.8163 | 0.78 | 0.78 |
78
- | 0.7077 | 2.9956 | 337 | 0.7802 | 0.8 | 0.8000 |
79
- | 0.1884 | 4.0 | 450 | 0.6157 | 0.87 | 0.87 |
80
- | 0.0209 | 4.9956 | 562 | 0.7885 | 0.84 | 0.8400 |
81
- | 0.117 | 6.0 | 675 | 0.6744 | 0.85 | 0.85 |
82
- | 0.0098 | 6.9956 | 787 | 0.6213 | 0.85 | 0.85 |
83
- | 0.0002 | 8.0 | 900 | 1.0599 | 0.82 | 0.82 |
84
- | 0.0001 | 8.9956 | 1012 | 0.7052 | 0.86 | 0.8600 |
85
- | 0.0001 | 10.0 | 1125 | 0.6891 | 0.85 | 0.85 |
86
- | 0.0001 | 10.9956 | 1237 | 0.6718 | 0.86 | 0.8600 |
87
- | 0.0001 | 12.0 | 1350 | 0.6712 | 0.85 | 0.85 |
88
- | 0.0001 | 12.9956 | 1462 | 0.6942 | 0.86 | 0.8600 |
89
- | 0.0001 | 14.0 | 1575 | 0.7002 | 0.86 | 0.8600 |
90
- | 0.0176 | 14.9956 | 1687 | 0.7053 | 0.86 | 0.8600 |
91
- | 0.0001 | 16.0 | 1800 | 0.7140 | 0.86 | 0.8600 |
92
- | 0.0001 | 16.9956 | 1912 | 0.7089 | 0.86 | 0.8600 |
93
- | 0.0 | 18.0 | 2025 | 0.7120 | 0.86 | 0.8600 |
94
- | 0.0628 | 18.9956 | 2137 | 0.7162 | 0.86 | 0.8600 |
95
- | 0.0037 | 19.9111 | 2240 | 0.7149 | 0.86 | 0.8600 |
96
 
97
 
98
  ### Framework versions
99
 
100
  - Transformers 4.42.4
101
- - Pytorch 2.3.1+cu121
102
  - Datasets 2.21.0
103
  - Tokenizers 0.19.1
 
23
  metrics:
24
  - name: Accuracy
25
  type: accuracy
26
+ value: 0.87
27
  - name: F1
28
  type: f1
29
+ value: 0.87
30
  ---
31
 
32
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
36
 
37
  This model is a fine-tuned version of [MIT/ast-finetuned-audioset-10-10-0.4593](https://huggingface.co/MIT/ast-finetuned-audioset-10-10-0.4593) on the GTZAN dataset.
38
  It achieves the following results on the evaluation set:
39
+ - Loss: 0.5658
40
+ - Accuracy: 0.87
41
+ - F1: 0.87
42
 
43
  ## Model description
44
 
 
61
  - train_batch_size: 2
62
  - eval_batch_size: 2
63
  - seed: 2024
64
+ - gradient_accumulation_steps: 8
65
+ - total_train_batch_size: 16
66
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
67
  - lr_scheduler_type: linear
68
  - lr_scheduler_warmup_ratio: 0.1
69
+ - num_epochs: 10
70
  - mixed_precision_training: Native AMP
71
 
72
  ### Training results
73
 
74
+ | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
75
+ |:-------------:|:------:|:----:|:---------------:|:--------:|:------:|
76
+ | 0.8357 | 0.9956 | 56 | 0.6582 | 0.82 | 0.82 |
77
+ | 0.4742 | 1.9911 | 112 | 0.6527 | 0.81 | 0.81 |
78
+ | 0.3344 | 2.9867 | 168 | 0.9048 | 0.76 | 0.76 |
79
+ | 0.0659 | 4.0 | 225 | 0.6998 | 0.84 | 0.8400 |
80
+ | 0.0966 | 4.9956 | 281 | 0.6737 | 0.83 | 0.83 |
81
+ | 0.0026 | 5.9911 | 337 | 0.5133 | 0.89 | 0.89 |
82
+ | 0.0038 | 6.9867 | 393 | 0.5704 | 0.86 | 0.8600 |
83
+ | 0.0005 | 8.0 | 450 | 0.5722 | 0.86 | 0.8600 |
84
+ | 0.0003 | 8.9956 | 506 | 0.5632 | 0.87 | 0.87 |
85
+ | 0.0003 | 9.9556 | 560 | 0.5658 | 0.87 | 0.87 |
 
 
 
 
 
 
 
 
 
 
86
 
87
 
88
  ### Framework versions
89
 
90
  - Transformers 4.42.4
91
+ - Pytorch 2.4.0+cu121
92
  - Datasets 2.21.0
93
  - Tokenizers 0.19.1
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:84d6d909b7239cd934a4caafdb079282c6dc019819cc1a8264706513cf4efd06
3
  size 344814656
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:693b8f1ddf6d4ea4efbcb7ff2e60a46edd67eb264433d7a798cff4f6e00eeda1
3
  size 344814656
runs/Aug27_09-10-11_75209667f075/events.out.tfevents.1724749825.75209667f075.596.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:192c251c128bc35fc3bcecd955173344cda774260c0b404ccb54d2751ee66e21
3
+ size 32884
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9bff867a7813abb44387806e4cea57940c830257824daf82a2f14b90ebf859d2
3
  size 5176
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0b0457c98a10fd7fea297df20ccd8ad719270517246fdf259a76de3e7033cc02
3
  size 5176