Whisper small pt jwlang - Michel Mesquita

This model is a fine-tuned version of openai/whisper-small on the jwlang 1.0 dataset. It achieves the following results on the evaluation set:

  • Loss: 0.6913
  • Wer: 23.2558

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 Wer
0.0053 14.0845 1000 0.6128 22.3990
0.0002 28.1690 2000 0.6528 22.6438
0.0001 42.2535 3000 0.6806 23.0110
0.0001 56.3380 4000 0.6913 23.2558

Framework versions

  • Transformers 4.41.2
  • Pytorch 2.3.0+cu121
  • Datasets 2.20.0
  • Tokenizers 0.19.1
Downloads last month
19
Safetensors
Model size
242M 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 M2LabOrg/whisper-small-pt-jwlang

Finetuned
(2102)
this model

Dataset used to train M2LabOrg/whisper-small-pt-jwlang

Evaluation results