README.md

metadata

language:
  - hi
license: apache-2.0
base_model: openai/whisper-medium
tags:
  - generated_from_trainer
datasets:
  - konnakol
metrics:
  - wer
model-index:
  - name: Whisper Hi - Gopika
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Konnakol
          type: konnakol
          args: 'config: hi, split: test'
        metrics:
          - name: Wer
            type: wer
            value: 0.01805869074492099

Whisper Hi - Gopika

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

• Loss: 0.1982
• Wer: 0.0181
• Cer: 0.0187

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-05
• train_batch_size: 32
• eval_batch_size: 16
• 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	Cer
2.7311	8.3333	50	0.7476	0.0339	0.0218
0.1743	16.6667	100	0.1654	0.0339	0.0305
0.0109	25.0	150	0.1474	0.0181	0.0187
0.0126	33.3333	200	0.1857	0.0316	0.0262
0.0122	41.6667	250	0.1125	0.0135	0.0143
0.0031	50.0	300	0.1541	0.0226	0.0224
0.0102	58.3333	350	0.1160	0.0226	0.0206
0.0211	66.6667	400	0.1925	0.0248	0.0224
0.0179	75.0	450	0.1253	0.0181	0.0187
0.0185	83.3333	500	0.1218	0.0271	0.0243
0.017	91.6667	550	0.1543	0.0316	0.0299
0.0077	100.0	600	0.1376	0.0226	0.0212
0.0033	108.3333	650	0.1555	0.0226	0.0206
0.0019	116.6667	700	0.1871	0.0203	0.0206
0.0012	125.0	750	0.1614	0.0226	0.0237
0.0	133.3333	800	0.1911	0.0181	0.0187
0.0	141.6667	850	0.1948	0.0181	0.0187
0.0	150.0	900	0.1968	0.0181	0.0187
0.0	158.3333	950	0.1977	0.0181	0.0187
0.0	166.6667	1000	0.1982	0.0181	0.0187

Framework versions

• Transformers 4.41.1
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1