nudity-pytorch

1. Nudity detection ML model in PyTorch
2. Mobile Nudity Detection ML Model in PyTorch iOS.

Nudity detection ML model in PyTorch

Welcome to the repository of our PyTorch model designed to detect nudity in images. This model is built on the philosophy that a simpler model with more data is always better than a larger model with less data. We present to you two differnt models including one for iOS devices.

Model Details

ResNet (Residual Network) is a deep neural network architecture introduced by Microsoft Research in 2015. It was primarily developed for image classification tasks and has since become a widely used and highly successful model in computer vision. The original paper titled "Deep Residual Learning for Image Recognition" is available at https://arxiv.org/abs/1512.03385.

The key innovation of ResNet is the introduction of residual connections, which allow the network to train deeper than previous architectures without suffering from vanishing gradient problems. These residual connections skip one or more layers in the network and add the output of those skipped layers to the input of the subsequent layers, creating a form of identity mapping that eases the training process.

This involves utilizing a pre-trained model and subsequently freezing the layers while adding an additional layer specifically designed for fine-tuning. In doing so, I have generated two distinct published models: one fine-tuned on a relatively smaller dataset (./trained_pytorch_model/resnet_finetuned_smaller_dataset.pth) and another trained on a more extensive dataset (WIP).

Model Performance

Our model has demonstrated high accuracy in its predictions. Here are some key performance metrics:

Confusion Matrix:
 [[3186    9]
 [  40 2969]]
Test Accuracy: 0.9921

Number of Positive Samples: 3076
Number of Negative Samples: 3177
Percentage of True Positives: 98.50%
Percentage of True Negatives: 99.31%

How to Use

Install Dependencies

Install the necessary dependencies using the following command:

pip install -r requirements.txt

Run the Model

Use the following command to run the model:

inference_model = torch.load('./trained_pytorch_model/inference_nudity_model.pth)
inference_model.eval()
model_prediction = inference_model(input)

Run the Mobile iOS device

Use the following command to run the model:

mobile_inference_model = torch.load('./trained_pytorch_model/mobile_version_resnet_finetuned_smaller_dataset.pth)
mobile_inference_model.eval()
model_prediction = mobile_inference_model(input)

Sample Outputs from the model

Model Prediction -> Contains Nudity

---

Model Prediction -> Image is Safe

---

Model Prediction -> Contains Nudity

---

Model Prediction -> Contains Nudity

---

Contributing

We welcome contributions to improve this model. Feel free to submit issues, fork the repository and send pull requests!

License

This project is licensed under the terms of the MIT license.

Contact

If you have any questions, feel free to reach out or submit an issue.

Pritish Yuvraj

Email: [email protected]

Linkedin -> https://www.linkedin.com/in/pritishyuvraj/

Thank you for visiting this repository, we hope you find it useful!