MTWD/whisper-small-test

Whisper Small fine tuned with comms

This model is a fine-tuned version of openai/whisper-small on the BrainHack ASR Test Two dataset. It achieves the following results on the evaluation set:

• Loss: 0.2141
• Wer: 0.0326

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: 4
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 10
• training_steps: 100
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0059	13.3333	20	0.1429	0.0380
0.0003	26.6667	40	0.2095	0.0380
0.0001	40.0	60	0.2166	0.0326
0.0001	53.3333	80	0.2154	0.0326
0.0001	66.6667	100	0.2141	0.0326

Framework versions

• Transformers 4.40.2
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1