You can deploy this model and start earning money today!

ArchiveofAkashic/
ResidualFasterRCNN

Name: ResidualFasterRCNN
Availability: InStock

Views

No views yet

Brain Tumor Detection with ResBlockRCNN (Faster R-CNN + Custom Backbone)

This repository contains code and resources for training and testing a brain tumor detection model using a custom ResVGG16 Lite backbone integrated with Faster R-CNN and RPN, implemented in PyTorch. The dataset format is COCO, and the workflow supports both training and inference/testing.

Requirements

Python 3.8+
PyTorch >= 1.10
torchvision

Install dependencies (if using virtual environment):

pip install torch torchvision

Dataset Preparation

Place your COCO-formatted dataset in Dataset/COCODataset/ with the following structure:

Dataset/COCODataset/
    train/
        ...images...
        _annotations.coco.json
    valid/
        ...images...
        _annotations.coco.json
    test/
        ...images...
        _annotations.coco.json

Ensure annotation files are in COCO format.
Loading the Model from Hugging Face You can load the model and its pre-trained weights directly using the transformers library.

Option A: Load with AutoModel (Recommended) If your model is compatible with Hugging Face's AutoModel (e.g., based on a standard architecture like Faster R-CNN):

Python

from transformers import AutoModel

model = AutoModel.from_pretrained("your-username/your-model-name")
model.eval()  # Set to evaluation mode

Option B: Load manual

Python

from huggingface_hub import hf_hub_download
import torch

# Download files
weights_path = hf_hub_download(
    repo_id="ArchiveofAkashic/ResidualFasterRCNN",
    filename="Best_RPNmodel.pth"  # Check the exact filename in the repo
)
config_path = hf_hub_download(
    repo_id="ArchiveofAkashic/ResidualFasterRCNN",
    filename="config.json"
)

Notes

Make sure your dataset paths and annotation files are correct.
For GPU acceleration, ensure CUDA is available and PyTorch is installed with CUDA support.
For custom datasets, adjust the number of classes and label mapping as needed.

Citation

If you use this code or model, please cite or acknowledge this repository.

For questions or issues, please open an issue or contact the maintainer.

ArchiveofAkashic/ResidualFasterRCNN