whisper-base-aug-11-may-lightning-v1

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

• Loss: 0.1011
• Wer: 83.8329

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: 3e-05
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• 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: 1000
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.7002	1.0	1414	0.1698	90.1418
0.1406	2.0	2828	0.1163	82.7225
0.0973	3.0	4242	0.1005	81.6232
0.0759	4.0	5656	0.0944	79.2710
0.0613	5.0	7070	0.0910	82.6157
0.0504	6.0	8484	0.0916	81.6877
0.0416	7.0	9898	0.0932	83.1141
0.0344	8.0	11312	0.0960	83.1408
0.0285	9.0	12726	0.0992	83.6482
0.0243	9.9933	14130	0.1011	83.8329

Framework versions

• Transformers 4.52.0.dev0
• Pytorch 2.7.0+cu128
• Datasets 3.5.1
• Tokenizers 0.21.1