Rziane's picture
Upload WhisperForConditionalGeneration
cb54854 verified
metadata
base_model: openai/whisper-medium
datasets:
  - AT_ENT
language:
  - aeb
library_name: transformers
license: apache-2.0
metrics:
  - wer
tags:
  - generated_from_trainer
model-index:
  - name: Whisper medium AT
    results:
      - task:
          type: automatic-speech-recognition
          name: Automatic Speech Recognition
        dataset:
          name: AT_ENT
          type: AT_ENT
          args: 'config: aeb, split: test'
        metrics:
          - type: wer
            value: 66.82967723906664
            name: Wer

Whisper medium AT

This model is a fine-tuned version of openai/whisper-medium on the AT_ENT dataset. It achieves the following results on the evaluation set:

  • Loss: 1.2100
  • Wer: 66.8297

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: 32
  • eval_batch_size: 16
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 500
  • num_epochs: 4
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
No log 1.0 111 1.0527 63.9374
No log 2.0 222 1.0961 65.8796
No log 3.0 333 1.1626 67.5283
No log 4.0 444 1.2100 66.8297

Framework versions

  • Transformers 4.45.1
  • Pytorch 2.4.1+cu121
  • Datasets 3.0.1
  • Tokenizers 0.20.0