whisper-a-nomimo-ls

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

• Loss: 0.0311
• Wer: 32.3302

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: 0.0004
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• 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: 132
• num_epochs: 11
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.9712	1.0	104	0.0786	25.0
3.0599	2.0	208	0.7840	3312.3457
0.3714	3.0	312	0.2694	98.3796
0.2625	4.0	416	0.2197	90.9722
0.2298	5.0	520	0.2035	84.9537
0.1929	6.0	624	0.1519	66.0494
0.1086	7.0	728	0.0641	41.6667
0.0401	8.0	832	0.0421	41.6667
0.0231	9.0	936	0.0324	34.8765
0.0119	10.0	1040	0.0310	34.7222
0.0057	10.8986	1133	0.0311	32.3302

Framework versions

• Transformers 4.47.0.dev0
• Pytorch 2.4.0
• Datasets 3.0.1
• Tokenizers 0.20.0