Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/alihassanml/fight-prediction-project
This project focuses on predicting fight actions using deep learning. We utilize models such as **ResNet152** and **EfficientNetB7** to classify different fight actions. The project predicts the following actions:
https://github.com/alihassanml/fight-prediction-project
cnn deep-learning efficientnet opencv rsnet50 streamlit
Last synced: about 1 month ago
JSON representation
This project focuses on predicting fight actions using deep learning. We utilize models such as **ResNet152** and **EfficientNetB7** to classify different fight actions. The project predicts the following actions:
- Host: GitHub
- URL: https://github.com/alihassanml/fight-prediction-project
- Owner: alihassanml
- License: mit
- Created: 2024-10-20T19:17:39.000Z (3 months ago)
- Default Branch: main
- Last Pushed: 2024-10-20T19:27:48.000Z (3 months ago)
- Last Synced: 2024-10-30T00:35:44.326Z (3 months ago)
- Topics: cnn, deep-learning, efficientnet, opencv, rsnet50, streamlit
- Language: Jupyter Notebook
- Homepage:
- Size: 2.05 MB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Fight Action Prediction
This project focuses on predicting fight actions using deep learning models. We classify various fight-related actions using models like **ResNet152** and **EfficientNetB7**. The actions being predicted include:
- **0**: Hit
- **1**: Kick
- **2**: Punch
- **3**: Push
- **4**: Ride Horse
- **5**: Shoot Gun
- **6**: Stand
- **7**: Wave
## Models
We trained and fine-tuned the following models:
- **ResNet152**: Fine-tuned on the action dataset.
- **EfficientNetB7**: Another deep learning model used for comparison.
## Dataset
The dataset for training includes labeled images corresponding to the actions mentioned above. Ensure your dataset follows the directory structure:
```
data/
│
├── train/
│ └── /
│ └── image1.jpg
├── test/
└── /
└── image2.jpg
```
Each class folder contains relevant images for that action.
## Installation
Clone the repository and install the dependencies:
```bash
git clone https://github.com/alihassanml/Fight-Prediction-Project.git
cd Fight-Prediction-Project
pip install -r requirements.txt
```
## Training
To train the models, run the following command:
```bash
python train.py
```
This will start training the model on the dataset located in the `data/train` directory.
## Streamlit Application
You can interact with the trained models using a web interface powered by Streamlit. To run the app:
```bash
streamlit run streamlit_app.py
```
Upload an image, select the model you want to use, and get predictions for the action being performed in the image.
## Results
After training, models will be saved as `.h5` files, and the model performance can be assessed using validation data and metrics such as accuracy, confusion matrix, and classification report.
## Fine-Tuning
Fine-tuning is done by unfreezing the last few layers of the pre-trained models and retraining them with a smaller learning rate to improve accuracy.
## License
This project is licensed under the MIT License.