metadata

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: 13.506122934252765

Whisper medium LV - Felikss Kleins

This model is a fine-tuned version of openai/whisper-medium on the Common Voice 17.0 dataset. It achieves the following results on the evaluation set:

Loss: 0.1848
Wer: 13.5061

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	0.04	200	0.8471	30.5453
1.8421	0.08	400	0.3571	27.1906
0.5851	0.12	600	0.3146	23.9508
0.4787	0.16	800	0.2913	23.2454
0.4845	0.2	1000	0.2736	20.9388
0.4845	0.24	1200	0.2495	19.1951
0.4004	0.28	1400	0.2416	18.4600
0.357	1.018	1600	0.2300	17.8576
0.3381	1.058	1800	0.2261	17.1125
0.3052	1.098	2000	0.2151	16.6013
0.3052	1.138	2200	0.2154	16.1673
0.2636	1.178	2400	0.2256	17.3107
0.2805	1.218	2600	0.2059	15.6482
0.2331	1.258	2800	0.2022	15.4599
0.2245	1.298	3000	0.1971	15.0953
0.2245	2.036	3200	0.1988	14.7604
0.2312	2.076	3400	0.1941	14.5623
0.2077	2.116	3600	0.1915	14.2175
0.1923	2.156	3800	0.1935	14.4018
0.2012	2.196	4000	0.1906	14.4850
0.2012	2.2360	4200	0.1880	13.9302
0.1694	2.276	4400	0.1860	14.1640
0.1548	3.014	4600	0.1855	13.6508
0.1628	3.054	4800	0.1850	13.5557
0.1529	3.094	5000	0.1848	13.5061

Framework versions

Transformers 4.45.0.dev0
Pytorch 2.0.1
Datasets 3.0.0
Tokenizers 0.19.1