Edit model card

whisper-large-v2-ca-10awr

This model is a fine-tuned version of openai/whisper-large-v2 on the cymen-arfor/15awr train[:66%] main dataset. It achieves the following results on the evaluation set:

  • Loss: 1.0685
  • Wer: 0.3837

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: 16
  • eval_batch_size: 16
  • 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: 500
  • training_steps: 5000
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
0.3063 4.7281 1000 0.6504 0.4067
0.0347 9.4563 2000 0.8513 0.3862
0.0141 14.1844 3000 0.9376 0.3877
0.0011 18.9125 4000 1.0259 0.3793
0.0007 23.6407 5000 1.0685 0.3837

Framework versions

  • Transformers 4.46.1
  • Pytorch 2.5.1+cu124
  • Datasets 3.0.2
  • Tokenizers 0.20.1
Downloads last month
23
Safetensors
Model size
1.54B params
Tensor type
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for cymen-arfor/whisper-large-v2-ca-10awr

Finetuned
(182)
this model