whisper-medium-ta-en

This model is a fine-tuned version of openai/whisper-medium on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.1186
• Bleu Score: 0.0

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: 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
• training_steps: 3000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu Score
0.1809	2.9412	250	0.0752	0.0
0.0082	5.8824	500	0.0883	0.0053
0.0008	8.8235	750	0.0959	0.0039
0.0003	11.7647	1000	0.1009	0.0045
0.0002	14.7059	1250	0.1052	0.0
0.0002	17.6471	1500	0.1089	0.0
0.0001	20.5882	1750	0.1119	0.0032
0.0001	23.5294	2000	0.1147	0.0
0.0001	26.4706	2250	0.1163	0.0
0.0001	29.4118	2500	0.1175	0.0
0.0001	32.3529	2750	0.1182	0.0
0.0001	35.2941	3000	0.1186	0.0

Framework versions

• Transformers 4.48.3
• Pytorch 2.6.0+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0