Commit
·
9e595ae
1
Parent(s):
9d47174
update model card README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,118 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- ru
|
| 4 |
+
license: apache-2.0
|
| 5 |
+
tags:
|
| 6 |
+
- hf-asr-leaderboard
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
datasets:
|
| 9 |
+
- mozilla-foundation/common_voice_11_0
|
| 10 |
+
metrics:
|
| 11 |
+
- wer
|
| 12 |
+
model-index:
|
| 13 |
+
- name: whisper-base-fine_tuned-ru
|
| 14 |
+
results:
|
| 15 |
+
- task:
|
| 16 |
+
name: Automatic Speech Recognition
|
| 17 |
+
type: automatic-speech-recognition
|
| 18 |
+
dataset:
|
| 19 |
+
name: common_voice_11_0
|
| 20 |
+
type: mozilla-foundation/common_voice_11_0
|
| 21 |
+
args: 'config: ru, split: test'
|
| 22 |
+
metrics:
|
| 23 |
+
- name: Wer
|
| 24 |
+
type: wer
|
| 25 |
+
value: 41.216909250757055
|
| 26 |
+
---
|
| 27 |
+
|
| 28 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 29 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 30 |
+
|
| 31 |
+
# whisper-base-fine_tuned-ru
|
| 32 |
+
|
| 33 |
+
This model is a fine-tuned version of [openai/whisper-base](https://huggingface.co/openai/whisper-base) on the common_voice_11_0 dataset.
|
| 34 |
+
It achieves the following results on the evaluation set:
|
| 35 |
+
- Loss: 0.4553
|
| 36 |
+
- Wer: 41.2169
|
| 37 |
+
|
| 38 |
+
## Model description
|
| 39 |
+
|
| 40 |
+
More information needed
|
| 41 |
+
|
| 42 |
+
## Intended uses & limitations
|
| 43 |
+
|
| 44 |
+
More information needed
|
| 45 |
+
|
| 46 |
+
## Training and evaluation data
|
| 47 |
+
|
| 48 |
+
More information needed
|
| 49 |
+
|
| 50 |
+
## Training procedure
|
| 51 |
+
|
| 52 |
+
### Training hyperparameters
|
| 53 |
+
|
| 54 |
+
The following hyperparameters were used during training:
|
| 55 |
+
- learning_rate: 1e-06
|
| 56 |
+
- train_batch_size: 4
|
| 57 |
+
- eval_batch_size: 4
|
| 58 |
+
- seed: 42
|
| 59 |
+
- gradient_accumulation_steps: 4
|
| 60 |
+
- total_train_batch_size: 16
|
| 61 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 62 |
+
- lr_scheduler_type: linear
|
| 63 |
+
- lr_scheduler_warmup_steps: 250
|
| 64 |
+
- training_steps: 20000
|
| 65 |
+
- mixed_precision_training: Native AMP
|
| 66 |
+
|
| 67 |
+
### Training results
|
| 68 |
+
|
| 69 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 70 |
+
|:-------------:|:-----:|:-----:|:---------------:|:-------:|
|
| 71 |
+
| 0.702 | 0.25 | 500 | 0.8245 | 71.6653 |
|
| 72 |
+
| 0.5699 | 0.49 | 1000 | 0.6640 | 55.7048 |
|
| 73 |
+
| 0.5261 | 0.74 | 1500 | 0.6127 | 50.6215 |
|
| 74 |
+
| 0.4997 | 0.98 | 2000 | 0.5834 | 47.4541 |
|
| 75 |
+
| 0.4681 | 1.23 | 2500 | 0.5638 | 46.6262 |
|
| 76 |
+
| 0.4651 | 1.48 | 3000 | 0.5497 | 47.5090 |
|
| 77 |
+
| 0.4637 | 1.72 | 3500 | 0.5379 | 46.5700 |
|
| 78 |
+
| 0.4185 | 1.97 | 4000 | 0.5274 | 45.3160 |
|
| 79 |
+
| 0.3856 | 2.22 | 4500 | 0.5205 | 45.5871 |
|
| 80 |
+
| 0.4078 | 2.46 | 5000 | 0.5122 | 45.7190 |
|
| 81 |
+
| 0.4132 | 2.71 | 5500 | 0.5066 | 45.5004 |
|
| 82 |
+
| 0.3914 | 2.96 | 6000 | 0.4998 | 47.0011 |
|
| 83 |
+
| 0.3822 | 3.2 | 6500 | 0.4959 | 44.9570 |
|
| 84 |
+
| 0.3596 | 3.45 | 7000 | 0.4916 | 45.5578 |
|
| 85 |
+
| 0.3877 | 3.69 | 7500 | 0.4870 | 45.2476 |
|
| 86 |
+
| 0.3687 | 3.94 | 8000 | 0.4832 | 45.2159 |
|
| 87 |
+
| 0.3514 | 4.19 | 8500 | 0.4809 | 46.0254 |
|
| 88 |
+
| 0.3202 | 4.43 | 9000 | 0.4779 | 48.1306 |
|
| 89 |
+
| 0.3229 | 4.68 | 9500 | 0.4751 | 45.5724 |
|
| 90 |
+
| 0.3285 | 4.93 | 10000 | 0.4717 | 45.9436 |
|
| 91 |
+
| 0.3286 | 5.17 | 10500 | 0.4705 | 45.0510 |
|
| 92 |
+
| 0.3294 | 5.42 | 11000 | 0.4689 | 47.2111 |
|
| 93 |
+
| 0.3384 | 5.66 | 11500 | 0.4666 | 47.3393 |
|
| 94 |
+
| 0.316 | 5.91 | 12000 | 0.4650 | 43.2536 |
|
| 95 |
+
| 0.2988 | 6.16 | 12500 | 0.4638 | 42.9789 |
|
| 96 |
+
| 0.3046 | 6.4 | 13000 | 0.4629 | 42.4331 |
|
| 97 |
+
| 0.2962 | 6.65 | 13500 | 0.4614 | 40.2437 |
|
| 98 |
+
| 0.3008 | 6.9 | 14000 | 0.4602 | 39.5734 |
|
| 99 |
+
| 0.2749 | 7.14 | 14500 | 0.4593 | 40.1497 |
|
| 100 |
+
| 0.3001 | 7.39 | 15000 | 0.4588 | 42.6248 |
|
| 101 |
+
| 0.3054 | 7.64 | 15500 | 0.4580 | 40.3707 |
|
| 102 |
+
| 0.2926 | 7.88 | 16000 | 0.4574 | 39.4232 |
|
| 103 |
+
| 0.2938 | 8.13 | 16500 | 0.4569 | 40.9532 |
|
| 104 |
+
| 0.3105 | 8.37 | 17000 | 0.4566 | 40.4379 |
|
| 105 |
+
| 0.2799 | 8.62 | 17500 | 0.4562 | 40.3622 |
|
| 106 |
+
| 0.2793 | 8.87 | 18000 | 0.4557 | 41.3451 |
|
| 107 |
+
| 0.2819 | 9.11 | 18500 | 0.4555 | 41.4184 |
|
| 108 |
+
| 0.2907 | 9.36 | 19000 | 0.4555 | 39.9348 |
|
| 109 |
+
| 0.3113 | 9.61 | 19500 | 0.4553 | 41.0289 |
|
| 110 |
+
| 0.2867 | 9.85 | 20000 | 0.4553 | 41.2169 |
|
| 111 |
+
|
| 112 |
+
|
| 113 |
+
### Framework versions
|
| 114 |
+
|
| 115 |
+
- Transformers 4.24.0
|
| 116 |
+
- Pytorch 1.13.1
|
| 117 |
+
- Datasets 2.7.1
|
| 118 |
+
- Tokenizers 0.13.1
|