metadata

language:
  - en
license: apache-2.0
base_model: openai/whisper-tiny
tags:
  - hf-asr-leaderboard
  - generated_from_trainer
datasets:
  - jpdiazpardo/guturalScream_metalVocals
model-index:
  - name: Whisper Tiny Metal - Juan Pablo Díaz
    results: []

Whisper Tiny Metal - Juan Pablo Díaz

This model is a fine-tuned version of openai/whisper-tiny on the Gutural Scream & Metal Vocals dataset.

Model description

The model is inteded for automatic speech recognition in gutural and scream voice. The model was trained on vocals preprocessed using Spleeter source separtion algorithm.

Intended uses & limitations

Check out a demo of the model in my 'Spaces' repository: jpdiazpardo/jpdiazpardo-whisper-tiny-metal

Load the dataset from huggingface in your notebook:

from transformers import WhisperForConditionalGeneration, WhisperProcessor

model = WhisperForConditionalGeneration.from_pretrained("jpdiazpardo/whisper-tiny-metal")
processor = WhisperProcessor.from_pretrained("jpdiazpardo/whisper-tiny-metal")

Training and evaluation data

jpdiazpardo/guturalScream_metalVocals

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: 1
training_steps: 2

Training results

Framework versions

Transformers 4.32.0
Pytorch 2.0.1+cu118
Datasets 2.14.4
Tokenizers 0.13.3