whisper-5hrs-random-seed42

This model is a fine-tuned version of openai/whisper-large-v2 on the JASMIN-CGN dataset. It achieves the following results on the evaluation set:

• Loss: 0.4224
• Wer: 19.5323

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: 48
• eval_batch_size: 32
• seed: 42
• optimizer: Use OptimizerNames.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: 11
• num_epochs: 3.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
1.0123	0.2973	11	1.1422	36.6894
0.8299	0.5946	22	0.8867	32.8446
0.6652	0.8919	33	0.6295	26.2690
0.6298	1.1892	44	0.5203	22.3806
0.5168	1.4865	55	0.4698	20.9011
0.4352	1.7838	66	0.4446	20.7770
0.4052	2.0811	77	0.4325	19.8443
0.4235	2.3784	88	0.4264	19.5021
0.4122	2.6757	99	0.4233	19.5256
0.4263	2.9730	110	0.4224	19.5323

Framework versions

• PEFT 0.16.0
• Transformers 4.57.3
• Pytorch 2.7.1+cu126
• Datasets 3.6.0
• Tokenizers 0.22.2