2026-02-25_02-15-51

This model is a fine-tuned version of openai/whisper-small.en on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2900
• Wer: 11.7695

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: 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: 500
• training_steps: 10000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
1.1276	0.1123	500	0.3356	13.7752
0.2792	0.2247	1000	0.3113	12.9679
0.2666	0.3370	1500	0.3035	12.6108
0.2616	0.4493	2000	0.2999	12.2032
0.2563	0.5617	2500	0.2986	13.2479
0.2501	0.6740	3000	0.2929	12.4547
0.2561	0.7863	3500	0.2910	11.9032
0.2497	0.8987	4000	0.2898	12.0265
0.2336	1.0110	4500	0.2877	11.4051
0.1795	1.1233	5000	0.2915	11.8667
0.1837	1.2357	5500	0.2909	11.7313
0.1763	1.3480	6000	0.2900	11.7695

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 3.6.0
• Tokenizers 0.21.1