Hugging Face's logo

spellingdragon
/
whisper-small-zh-foriegn

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
Chinese
whisper
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
spellingdragon commited on
Commit
90a9ea3
·
1 Parent(s): ae1534c

End of training

Browse files
Files changed (6) hide show
  1. README.md +24 -15
  2. adapter_config.json +30 -0
  3. adapter_model.safetensors +3 -0
  4. runs/Dec17_10-38-30_1dee50c2882e/events.out.tfevents.1702809511.1dee50c2882e.344.0 +3 -0
  5. runs/Dec17_11-22-10_1dee50c2882e/events.out.tfevents.1702812131.1dee50c2882e.344.1 +3 -0
  6. training_args.bin +1 -1
README.md CHANGED
@@ -2,7 +2,7 @@
2
  language:
3
  - zh
4
  license: apache-2.0
5
- base_model: openai/whisper-large-v3
6
  tags:
7
  - generated_from_trainer
8
  datasets:
@@ -17,16 +17,9 @@ should probably proofread and complete it, then remove this comment. -->
17
 
18
  # Whisper small zh - foriegn
19
 
20
- This model is a fine-tuned version of [openai/whisper-large-v3](https://huggingface.co/openai/whisper-large-v3) on the Common Voice 9 dataset.
21
  It achieves the following results on the evaluation set:
22
- - eval_loss: 0.1889
23
- - eval_wer_ortho: 39.8441
24
- - eval_wer: 39.8747
25
- - eval_runtime: 1418.0546
26
- - eval_samples_per_second: 3.257
27
- - eval_steps_per_second: 0.204
28
- - epoch: 3.63
29
- - step: 2500
30
 
31
  ## Model description
32
 
@@ -45,19 +38,35 @@ More information needed
45
  ### Training hyperparameters
46
 
47
  The following hyperparameters were used during training:
48
- - learning_rate: 1e-05
49
- - train_batch_size: 16
50
- - eval_batch_size: 16
51
  - seed: 42
 
 
52
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
53
  - lr_scheduler_type: linear
54
  - lr_scheduler_warmup_steps: 50
55
- - training_steps: 4000
56
  - mixed_precision_training: Native AMP
57
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
58
  ### Framework versions
59
 
60
  - Transformers 4.35.2
61
- - Pytorch 2.1.0+cu118
62
  - Datasets 2.15.0
63
  - Tokenizers 0.15.0
 
2
  language:
3
  - zh
4
  license: apache-2.0
5
+ base_model: openai/whisper-small
6
  tags:
7
  - generated_from_trainer
8
  datasets:
 
17
 
18
  # Whisper small zh - foriegn
19
 
20
+ This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 9 dataset.
21
  It achieves the following results on the evaluation set:
22
+ - Loss: 0.2235
 
 
 
 
 
 
 
23
 
24
  ## Model description
25
 
 
38
  ### Training hyperparameters
39
 
40
  The following hyperparameters were used during training:
41
+ - learning_rate: 0.001
42
+ - train_batch_size: 2
43
+ - eval_batch_size: 8
44
  - seed: 42
45
+ - gradient_accumulation_steps: 4
46
+ - total_train_batch_size: 8
47
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
48
  - lr_scheduler_type: linear
49
  - lr_scheduler_warmup_steps: 50
50
+ - num_epochs: 3
51
  - mixed_precision_training: Native AMP
52
 
53
+ ### Training results
54
+
55
+ | Training Loss | Epoch | Step | Validation Loss |
56
+ |:-------------:|:-----:|:----:|:---------------:|
57
+ | 0.2446 | 0.36 | 500 | 0.3424 |
58
+ | 0.3316 | 0.73 | 1000 | 0.3260 |
59
+ | 0.2777 | 1.09 | 1500 | 0.2939 |
60
+ | 0.1808 | 1.45 | 2000 | 0.2944 |
61
+ | 0.1822 | 1.82 | 2500 | 0.2589 |
62
+ | 0.1206 | 2.18 | 3000 | 0.2468 |
63
+ | 0.0715 | 2.54 | 3500 | 0.2319 |
64
+ | 0.0642 | 2.91 | 4000 | 0.2235 |
65
+
66
+
67
  ### Framework versions
68
 
69
  - Transformers 4.35.2
70
+ - Pytorch 2.1.0+cu121
71
  - Datasets 2.15.0
72
  - Tokenizers 0.15.0
adapter_config.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": {
4
+ "base_model_class": "WhisperForConditionalGeneration",
5
+ "parent_library": "transformers.models.whisper.modeling_whisper"
6
+ },
7
+ "base_model_name_or_path": "openai/whisper-small",
8
+ "bias": "none",
9
+ "fan_in_fan_out": false,
10
+ "inference_mode": true,
11
+ "init_lora_weights": true,
12
+ "layers_pattern": null,
13
+ "layers_to_transform": null,
14
+ "loftq_config": {},
15
+ "lora_alpha": 64,
16
+ "lora_dropout": 0.05,
17
+ "megatron_config": null,
18
+ "megatron_core": "megatron.core",
19
+ "modules_to_save": null,
20
+ "peft_type": "LORA",
21
+ "r": 32,
22
+ "rank_pattern": {},
23
+ "revision": null,
24
+ "target_modules": [
25
+ "v_proj",
26
+ "q_proj"
27
+ ],
28
+ "task_type": null,
29
+ "use_rslora": false
30
+ }
adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:92d4f4b3c69b7d1c965a6754616c1d598bd40d53f71b24be0b07b044a34e64f4
3
+ size 14176064
runs/Dec17_10-38-30_1dee50c2882e/events.out.tfevents.1702809511.1dee50c2882e.344.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8c97b5a146dc624cd510b34a9ccbf293a7ea0107050de64f72818fa1cfd18ffe
3
+ size 6448
runs/Dec17_11-22-10_1dee50c2882e/events.out.tfevents.1702812131.1dee50c2882e.344.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1155fa6d05880a025e650ef4247e6244d8f88e760fbb6f8639ca2cf22493ea3b
3
+ size 9798
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e1534174fcd95c994aab11c793e457fb7d21cc22944ea2cc2221a22d101a3e63
3
  size 4728
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0d713f347acb1e26d4873cf3be75889b99d2e72e6e28b14831209a691ac58e40
3
  size 4728