piyushmaharana/outcomes-whisper-tiny-v1.2

OutcomesAI-Whisper-tiny-v1.2

This model is a fine-tuned version of piyushmaharana/outcomes-whisper-tiny-v1.1 on the big-transcript-pronounce dataset. It achieves the following results on the evaluation set:

• Loss: 0.0256
• Wer: 1.7139

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: 5e-06
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• 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: 500
• training_steps: 3000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.112	2.1645	500	0.0645	5.5784
0.013	4.3290	1000	0.0310	2.3237
0.0032	6.4935	1500	0.0264	1.7798
0.0012	8.6580	2000	0.0260	1.7304
0.0007	10.8225	2500	0.0257	1.7963
0.0006	12.9870	3000	0.0256	1.7139

Framework versions

• Transformers 4.44.0
• Pytorch 2.4.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1