whisper-large-ver2

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

  • Loss: 0.4756
  • Cer: 11.2426

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: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 500
  • training_steps: 4000
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Cer
0.0327 5.59 1000 0.3779 14.1439
0.004 11.17 2000 0.4122 13.6476
0.0005 16.76 3000 0.4584 11.2044
0.0004 22.35 4000 0.4756 11.2426

Framework versions

  • Transformers 4.39.0.dev0
  • Pytorch 2.0.0+cu118
  • Datasets 2.18.0
  • Tokenizers 0.15.2
Downloads last month
19
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 unanam/whisper-large-ver2

Finetuned
(191)
this model