https://github.com/yidadaa/pytorch-video-classification

Make video classification on UCF101 using CNN and RNN based on Pytorch framework.
https://github.com/yidadaa/pytorch-video-classification

pytorch video-classification

Last synced: about 1 month ago
JSON representation

Make video classification on UCF101 using CNN and RNN based on Pytorch framework.

• Host: GitHub
• URL: https://github.com/yidadaa/pytorch-video-classification
• Owner: Yidadaa
• Created: 2019-06-05T12:14:08.000Z (over 5 years ago)
• Default Branch: master
• Last Pushed: 2023-12-04T08:18:59.000Z (about 1 year ago)
• Last Synced: 2025-01-03T04:08:44.813Z (about 2 months ago)
• Topics: pytorch, video-classification
• Language: Python
• Size: 48.8 KB
• Stars: 62
• Watchers: 5
• Forks: 29
• Open Issues: 1
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Pytorch-Video-Classification

Make video classification on UCF101 using CNN and RNN with Pytorch framework.

# Environments

```bash

# 1. torch >= 1.0

conda create -n crnn

source activate crnn # or `conda activate crnn`

# GPU version

conda install pytorch torchvision cudatoolkit=9.0 -c pytorch

# CPU version

conda install pytorch-cpu torchvision-cpu -c pytorch

# 2. pip dependencies

pip install pandas scikit-learn tqdm opencv-python

# 3. prepare datasets

cd ./Pytorch-Video-Classification # go to the root dir of the code

cp -r path/to/your/UCF ./data # copy UCF dataset to data dir

cd ./data && python make_train_test.py # preprocess the dataset

# 4. train your network on UCF-101

python train.py

# (optional)5. restore from checkpoints

python train.py -r path/to/checkpoints/file

```

To know more about the usage of scripts, run the following commands:

```bash

python train.py -h

python make_train_test.py -h

```