Rycen7822 huggingface-vision-trainer
When to Use
Use for image classification or vision model training pipelines, dataset transforms, evaluation metrics, and checkpoint publishing.
Do Not Use When
Do not use for text-only LLM fine-tuning or Gradio UI-only changes.
Required Inputs
Image dataset, model checkpoint, transforms, metrics, hardware constraints, and output repo or checkpoint path.
Workflow
- Load and inspect image dataset splits.
- Define transforms and trainer config.
- Run a small smoke train/eval.
- Save and optionally push the checkpoint.