whisper-ft-large-1000-f

This model is a fine-tuned version of openai/whisper-large-v3-turbo on the None dataset. It achieves the following results on the evaluation set:

• Loss: 3.9591

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: 1
• eval_batch_size: 1
• seed: 42
• optimizer: Use OptimizerNames.PAGED_ADAMW_8BIT 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: 10
• training_steps: 40
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
4.6008	0.0029	2	6.9022
4.3327	0.0057	4	6.9022
4.6788	0.0086	6	6.7334
4.1284	0.0114	8	5.6978
2.5702	0.0143	10	4.6321
1.5963	0.0171	12	4.3133
1.3669	0.02	14	4.2577
1.1867	0.0229	16	4.3040
1.7891	0.0257	18	4.3839
1.14	0.0286	20	5.0947
0.7743	0.0314	22	4.4104
0.965	0.0343	24	4.0427
0.8264	0.0371	26	3.9472
0.7508	0.04	28	3.9840
0.4857	0.0429	30	4.0175
0.609	0.0457	32	4.2039
1.2697	0.0486	34	4.1989
0.6072	0.0514	36	4.0805
0.6064	0.0543	38	3.9941
0.7662	0.0571	40	3.9591

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0