README.md

---
language:
- ru
base_model: openai/whisper-base
tags:
- generated_from_trainer
datasets:
- aangry-mouse/stepik_ml_ru
metrics:
- wer
model-index:
- name: Whisper Base Ml Ru
  results:
  - task:
      name: Automatic Speech Recognition
      type: automatic-speech-recognition
    dataset:
      name: "ML \u0434\u0430\u0442\u0430\u0441\u0435\u0442"
      type: aangry-mouse/stepik_ml_ru
      args: 'config: ru, split: test'
    metrics:
    - name: Wer
      type: wer
      value: 35.272870043188064
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# Whisper Base Ml Ru

This model is a fine-tuned version of [openai/whisper-base](https://huggingface.co/openai/whisper-base) on the ML датасет dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4705
- Wer: 35.2729

## Model description

More information needed

## Intended uses & limitations

More information needed

## Training and evaluation data

More information needed

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 1000
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch  | Step | Validation Loss | Wer     |
|:-------------:|:------:|:----:|:---------------:|:-------:|
| 0.6752        | 0.6649 | 250  | 0.6513          | 39.3404 |
| 0.4415        | 1.3298 | 500  | 0.5277          | 38.4923 |
| 0.4037        | 1.9947 | 750  | 0.4766          | 35.0766 |
| 0.2825        | 2.6596 | 1000 | 0.4705          | 35.2729 |


### Framework versions

- Transformers 4.41.0
- Pytorch 2.0.1+cu118
- Datasets 2.19.1
- Tokenizers 0.19.1