Thecoder3281f/whisper-small-hi-commonvoice17-1000

Whisper Small English (1000 steps) - Jarrett Er

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

• Loss: 0.2214
• Wer: 10.9983

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: 0.0001
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• training_steps: 1000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.2461	0.1	100	0.2450	11.8146
0.2253	0.2	200	0.2343	11.2599
0.2663	0.3	300	0.2287	11.2495
0.252	0.4	400	0.2241	11.0506
0.2301	0.5	500	0.2245	11.4064
0.1985	1.037	600	0.2208	11.0297
0.1768	1.137	700	0.2230	11.2181
0.1107	1.237	800	0.2226	11.0820
0.0819	1.337	900	0.2228	11.1134
0.185	1.437	1000	0.2214	10.9983

Framework versions

• PEFT 0.14.0
• Transformers 4.48.0.dev0
• Pytorch 2.5.1+cu124
• Datasets 3.2.1.dev0
• Tokenizers 0.21.0