arun100/whisper-base-hi-1

Whisper Base Hindi

This model is a fine-tuned version of openai/whisper-base on the mozilla-foundation/common_voice_16_0 hi dataset. It achieves the following results on the evaluation set:

• Loss: 0.4926
• Wer: 27.4342

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 2000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.563	2.02	200	0.6270	38.2146
0.3107	5.01	400	0.4695	30.0641
0.1535	7.03	600	0.4548	27.7139
0.0841	10.02	800	0.4926	27.4342
0.0357	13.01	1000	0.5585	28.1772
0.0152	15.03	1200	0.6247	28.0687
0.0063	18.02	1400	0.6796	28.1856
0.0036	21.0	1600	0.7097	28.2670
0.0029	23.03	1800	0.7270	28.1960
0.0024	26.01	2000	0.7336	28.2649

Framework versions

• Transformers 4.37.0.dev0
• Pytorch 2.1.2+cu121
• Datasets 2.16.2.dev0
• Tokenizers 0.15.0