metadata

library_name: peft
language:
  - en
license: apache-2.0
base_model: openai/whisper-small
tags:
  - generated_from_trainer
datasets:
  - mesolitica/IMDA-TTS
metrics:
  - wer
model-index:
  - name: Whisper Small NSC small (500 steps) - Jarrett Er
    results:
      - task:
          type: automatic-speech-recognition
          name: Automatic Speech Recognition
        dataset:
          name: NSC Small section
          type: mesolitica/IMDA-TTS
          split: None
          args: 'config: en, split: train'
        metrics:
          - type: wer
            value: 3.0164184803360063
            name: Wer

Whisper Small NSC small (500 steps) - Jarrett Er

This model is a fine-tuned version of Thecoder3281f/whisper-small-hi-commonvoice17-1000 on the NSC Small section dataset. It achieves the following results on the evaluation set:

Loss: 0.0777
Wer: 3.0164

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: 0.0001
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: 50
training_steps: 500
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0822	0.8850	100	0.0686	3.0164
0.0585	1.7699	200	0.0700	3.0928
0.0317	2.6549	300	0.0726	3.0546
0.0184	3.5398	400	0.0781	3.2455
0.0194	4.4248	500	0.0777	3.0164

Framework versions

PEFT 0.14.0
Transformers 4.45.2
Pytorch 2.5.1+cu124
Datasets 3.2.1.dev0
Tokenizers 0.20.3