501Good/whisper-small-tt

Whisper Small Tatar - Kirill Milintsevich

This model is a fine-tuned version of openai/whisper-small on the Common Voice 11.0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4745
• Wer: 29.1407

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: 64
• eval_batch_size: 8
• 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: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0415	4.98	1000	0.3252	29.9512
0.0041	9.95	2000	0.3982	29.4805
0.0007	14.93	3000	0.4457	29.1513
0.0003	19.9	4000	0.4665	29.0912
0.0002	24.88	5000	0.4745	29.1407