Hugging Face's logo

miosipof
/
whisper-small-ft-balbus-sep28k-multiclass_v3

Audio Classification
Transformers
Safetensors
whisper
Generated from Trainer
Eval Results
Model card Files Community
miosipof commited on
Commit
d3f9299
·
verified ·
1 Parent(s): 039bd50

End of training

Browse files
Files changed (1) hide show
  1. README.md +25 -25
README.md CHANGED
@@ -23,16 +23,16 @@ model-index:
23
  metrics:
24
  - name: Accuracy
25
  type: accuracy
26
- value: 0.37653846153846154
27
  - name: Precision
28
  type: precision
29
- value: 0.1255128205128205
30
  - name: Recall
31
  type: recall
32
- value: 0.3333333333333333
33
  - name: F1
34
  type: f1
35
- value: 0.18236006333240196
36
  ---
37
 
38
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -42,11 +42,11 @@ should probably proofread and complete it, then remove this comment. -->
42
 
43
  This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Apple dataset dataset.
44
  It achieves the following results on the evaluation set:
45
- - Loss: 0.9882
46
- - Accuracy: 0.3765
47
- - Precision: 0.1255
48
- - Recall: 0.3333
49
- - F1: 0.1824
50
  - Roc-auc: None
51
 
52
  ## Model description
@@ -66,12 +66,12 @@ More information needed
66
  ### Training hyperparameters
67
 
68
  The following hyperparameters were used during training:
69
- - learning_rate: 0.001
70
- - train_batch_size: 8
71
- - eval_batch_size: 4
72
  - seed: 42
73
  - gradient_accumulation_steps: 4
74
- - total_train_batch_size: 32
75
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
76
  - lr_scheduler_type: linear
77
  - lr_scheduler_warmup_ratio: 0.5
@@ -82,18 +82,18 @@ The following hyperparameters were used during training:
82
 
83
  | Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 | Roc-auc |
84
  |:-------------:|:------:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|:-------:|
85
- | 0.8577 | 0.1368 | 100 | 0.8818 | 0.5219 | 0.5566 | 0.5095 | 0.4750 | None |
86
- | 1.0396 | 0.2735 | 200 | 1.0092 | 0.3335 | 0.1112 | 0.3333 | 0.1667 | None |
87
- | 1.1409 | 0.4103 | 300 | 1.0101 | 0.3746 | 0.2324 | 0.3353 | 0.2285 | None |
88
- | 1.0003 | 0.5470 | 400 | 1.0013 | 0.3765 | 0.1255 | 0.3333 | 0.1824 | None |
89
- | 1.0057 | 0.6838 | 500 | 1.0001 | 0.3765 | 0.1255 | 0.3333 | 0.1824 | None |
90
- | 0.993 | 0.8205 | 600 | 1.0461 | 0.3765 | 0.1255 | 0.3333 | 0.1824 | None |
91
- | 1.0017 | 0.9573 | 700 | 0.9998 | 0.3404 | 0.2501 | 0.3351 | 0.2226 | None |
92
- | 0.9992 | 1.0940 | 800 | 0.9927 | 0.3715 | 0.2364 | 0.3344 | 0.2418 | None |
93
- | 0.9805 | 1.2308 | 900 | 0.9965 | 0.3765 | 0.1255 | 0.3333 | 0.1824 | None |
94
- | 0.9934 | 1.3675 | 1000 | 0.9944 | 0.3804 | 0.2501 | 0.3499 | 0.2848 | None |
95
- | 0.9961 | 1.5043 | 1100 | 0.9899 | 0.3792 | 0.3125 | 0.3363 | 0.1928 | None |
96
- | 0.9878 | 1.6410 | 1200 | 0.9882 | 0.3765 | 0.1255 | 0.3333 | 0.1824 | None |
97
 
98
 
99
  ### Framework versions
 
23
  metrics:
24
  - name: Accuracy
25
  type: accuracy
26
+ value: 0.6703846153846154
27
  - name: Precision
28
  type: precision
29
+ value: 0.6905353453494015
30
  - name: Recall
31
  type: recall
32
+ value: 0.6572632164958351
33
  - name: F1
34
  type: f1
35
+ value: 0.6613993030224551
36
  ---
37
 
38
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
42
 
43
  This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Apple dataset dataset.
44
  It achieves the following results on the evaluation set:
45
+ - Loss: 0.5221
46
+ - Accuracy: 0.6704
47
+ - Precision: 0.6905
48
+ - Recall: 0.6573
49
+ - F1: 0.6614
50
  - Roc-auc: None
51
 
52
  ## Model description
 
66
  ### Training hyperparameters
67
 
68
  The following hyperparameters were used during training:
69
+ - learning_rate: 1e-05
70
+ - train_batch_size: 12
71
+ - eval_batch_size: 6
72
  - seed: 42
73
  - gradient_accumulation_steps: 4
74
+ - total_train_batch_size: 48
75
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
76
  - lr_scheduler_type: linear
77
  - lr_scheduler_warmup_ratio: 0.5
 
82
 
83
  | Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 | Roc-auc |
84
  |:-------------:|:------:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|:-------:|
85
+ | 0.7619 | 0.2051 | 100 | 0.7516 | 0.4035 | 0.4054 | 0.3432 | 0.2112 | None |
86
+ | 0.6854 | 0.4103 | 200 | 0.6178 | 0.5138 | 0.6335 | 0.4817 | 0.4125 | None |
87
+ | 0.5066 | 0.6154 | 300 | 0.5008 | 0.6231 | 0.6986 | 0.5986 | 0.5812 | None |
88
+ | 0.4635 | 0.8205 | 400 | 0.4840 | 0.6331 | 0.6968 | 0.6098 | 0.5959 | None |
89
+ | 0.4525 | 1.0256 | 500 | 0.5023 | 0.6469 | 0.6885 | 0.6255 | 0.6287 | None |
90
+ | 0.4164 | 1.2308 | 600 | 0.5023 | 0.6719 | 0.6784 | 0.6697 | 0.6730 | None |
91
+ | 0.3798 | 1.4359 | 700 | 0.4758 | 0.6412 | 0.6931 | 0.6192 | 0.6161 | None |
92
+ | 0.4174 | 1.6410 | 800 | 0.4708 | 0.66 | 0.6998 | 0.6395 | 0.6410 | None |
93
+ | 0.4212 | 1.8462 | 900 | 0.4717 | 0.6796 | 0.6951 | 0.6696 | 0.6758 | None |
94
+ | 0.3413 | 2.0513 | 1000 | 0.5100 | 0.6769 | 0.6904 | 0.6666 | 0.6699 | None |
95
+ | 0.2757 | 2.2564 | 1100 | 0.5297 | 0.6735 | 0.6892 | 0.6630 | 0.6682 | None |
96
+ | 0.2894 | 2.4615 | 1200 | 0.5221 | 0.6704 | 0.6905 | 0.6573 | 0.6614 | None |
97
 
98
 
99
  ### Framework versions