---
library_name: transformers
language:
- lv
license: apache-2.0
base_model: openai/whisper-medium
tags:
- hf-asr-leaderboard
- generated_from_trainer
datasets:
- mozilla-foundation/common_voice_17_0
metrics:
- wer
model-index:
- name: Whisper medium LV - Felikss Kleins
  results:
  - task:
      name: Automatic Speech Recognition
      type: automatic-speech-recognition
    dataset:
      name: Common Voice 17.0
      type: mozilla-foundation/common_voice_17_0
      config: lv
      split: None
      args: 'config: lv, split: test'
    metrics:
    - name: Wer
      type: wer
      value: 67.95180722891565
---

<!-- 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 medium LV - Felikss Kleins

This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the Common Voice 17.0 dataset.
It achieves the following results on the evaluation set:
- Loss: 1.1313
- Wer: 67.9518

## 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: 5e-06
- train_batch_size: 6
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 12
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 5000
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch    | Step | Validation Loss | Wer     |
|:-------------:|:--------:|:----:|:---------------:|:-------:|
| No log        | 23.0009  | 200  | 0.9920          | 47.2289 |
| 1.4829        | 47.0001  | 400  | 0.7982          | 48.9157 |
| 0.046         | 70.001   | 600  | 0.9576          | 51.3253 |
| 0.011         | 94.0002  | 800  | 0.9129          | 49.3976 |
| 0.0057        | 117.0011 | 1000 | 0.9789          | 51.3253 |
| 0.0057        | 141.0003 | 1200 | 1.0248          | 51.8072 |
| 0.005         | 164.0012 | 1400 | 1.0504          | 53.4940 |
| 0.0021        | 188.0004 | 1600 | 1.0447          | 67.9518 |
| 0.0015        | 211.0013 | 1800 | 1.0433          | 73.2530 |
| 0.0012        | 235.0005 | 2000 | 1.0646          | 55.1807 |
| 0.0012        | 258.0014 | 2200 | 1.1244          | 53.7349 |
| 0.0007        | 282.0006 | 2400 | 1.1156          | 59.5181 |
| 0.0006        | 305.0015 | 2600 | 1.1081          | 58.5542 |
| 0.0009        | 329.0007 | 2800 | 1.0342          | 54.4578 |
| 0.0006        | 352.0016 | 3000 | 1.0215          | 50.8434 |
| 0.0006        | 376.0008 | 3200 | 1.0619          | 56.6265 |
| 0.0004        | 399.0017 | 3400 | 1.1083          | 55.4217 |
| 0.0003        | 423.0009 | 3600 | 1.0970          | 56.8675 |
| 0.0006        | 447.0001 | 3800 | 1.1047          | 59.0361 |
| 0.0003        | 470.001  | 4000 | 1.1033          | 56.1446 |
| 0.0003        | 494.0002 | 4200 | 1.1003          | 57.5904 |
| 0.0002        | 517.0011 | 4400 | 1.1133          | 68.4337 |
| 0.0002        | 541.0003 | 4600 | 1.1146          | 69.3976 |
| 0.0001        | 564.0012 | 4800 | 1.1267          | 69.8795 |
| 0.0001        | 588.0004 | 5000 | 1.1313          | 67.9518 |


### Framework versions

- Transformers 4.45.0.dev0
- Pytorch 2.0.1
- Datasets 3.0.0
- Tokenizers 0.19.1