steja/whisper-small-marathi

Whisper Small Marathi

This model is a fine-tuned version of openai/whisper-small on the mozilla-foundation/common_voice_11_0 mr dataset. It achieves the following results on the evaluation set:

• Loss: 0.6231
• Wer: 18.5782

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: 16
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 64
• total_eval_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: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0027	18.87	1000	0.5090	20.0322
0.0007	37.74	2000	0.5729	19.0012
0.0	56.6	3000	0.6231	18.5782
0.0	75.47	4000	0.6435	18.6238
0.0	94.34	5000	0.6520	18.6046

Framework versions

• Transformers 4.25.1
• Pytorch 1.13.0+cu117
• Datasets 2.7.1
• Tokenizers 0.13.2