Hugging Face's logo

davidilag
/
whisper-tiny-no-fo-100h-5k-steps_v2

Automatic Speech Recognition
Transformers
Safetensors
whisper
Generated from Trainer
Inference Endpoints
Model card Files Community
whisper-tiny-no-fo-100h-5k-steps_v2 / README.md
davidilag's picture
davidilag
End of training
9ec8aef verified
preview code
|
raw
history blame contribute delete
1.76 kB
metadata
library_name: transformers
license: apache-2.0
base_model: NbAiLab/nb-whisper-tiny
tags:
  - generated_from_trainer
metrics:
  - wer
model-index:
  - name: whisper-tiny-no-fo-100h-5k-steps_v2
    results: []

whisper-tiny-no-fo-100h-5k-steps_v2

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

  • Loss: 0.5488
  • Wer: 56.4778

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: 5000
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
1.08 0.2320 1000 1.0274 74.5369
0.7205 0.4640 2000 0.7131 63.4588
0.6264 0.6961 3000 0.6096 58.1286
0.624 0.9281 4000 0.5642 56.2462
0.5163 1.1601 5000 0.5488 56.4778

Framework versions

  • Transformers 4.44.2
  • Pytorch 2.4.0+cu121
  • Datasets 3.0.0
  • Tokenizers 0.19.1