https://github.com/sanket-poojary-03/fine-tuning-vivit

Python script to fine tune Open source Video Vision Transformer (ViVit) using HuggingFace Trainer Library
https://github.com/sanket-poojary-03/fine-tuning-vivit

deep-learning fine-tuning huggingface huggingface-transformers pretrained-models python pytorch video-classification vision-transformer vivit

Last synced: 2 months ago
JSON representation

Python script to fine tune Open source Video Vision Transformer (ViVit) using HuggingFace Trainer Library

• Host: GitHub
• URL: https://github.com/sanket-poojary-03/fine-tuning-vivit
• Owner: sanket-poojary-03
• Created: 2024-07-26T14:32:07.000Z (6 months ago)
• Default Branch: main
• Last Pushed: 2024-08-01T04:55:06.000Z (5 months ago)
• Last Synced: 2024-10-10T19:14:57.220Z (3 months ago)
• Topics: deep-learning, fine-tuning, huggingface, huggingface-transformers, pretrained-models, python, pytorch, video-classification, vision-transformer, vivit
• Language: Python
• Homepage:
• Size: 24.4 KB
• Stars: 5
• Watchers: 5
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Fine-Tuning ViVit with HuggingFace Trainer

This repository contains a Python script to fine-tune the open-source Video Vision Transformer (ViVit) model using the HuggingFace Trainer Library. The model has been configured for 10 classes.

## Introduction

The Video Vision Transformer (ViVit) is a state-of-the-art model for video understanding tasks. This repository provides a script to fine-tune the ViVit model on your custom dataset using the HuggingFace Trainer Library. The model is pre-configured to classify videos into 10 different classes.

## Installation

To get started, clone this repository and install the required dependencies:

```bash

pip install -r requirements.txt

```

## Dataset Preparation

Prepare your dataset in the following format:

```python

DatasetDict({

    train: Dataset({

        features: ['labels', 'pixel_values'],

        num_rows: 36

    })

    test: Dataset({

        features: ['labels', 'pixel_values'],

        num_rows: 4

    })

})

```