brats20-3d-unet-pytorch is an open-weight model from Sagar Kumar. Its published files total 19.3 MB.
This repository contains the weights for a 3D U-Net model trained from scratch on the BraTS 2020 dataset for volumetric medical image segmentation.
Model Card
This repository contains the weights for a 3D U-Net model trained from scratch on the BraTS 2020 dataset for volumetric medical image segmentation. The model identifies and segments complex brain tumor sub-regions using multi-modal MRI scans (T1, T1ce, T2, FLAIR). The training pipeline was engineered using PyTorch and MONAI, featuring Automatic Mixed Precision (AMP) and overlapping sliding-window inference optimized for 16GB GPU constraints. (Note: Upload the Multi-Planar 3D reconstruction and Class-Wise Error Map images you generated in Kaggle directly here using the image upload button in the Hugging Face editor.)
Excerpt from the card by Sagar Kumar.
Identity and Version
- Repository
- sagarguptaiitp/brats20-3d-unet-pytorch
- Publisher
- Sagar Kumar
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- Not stated by the source
- Languages
- Not stated by the source
- Revision
- db52ec9bff20bb15ce4e8f22122db375e797eb12
- First published
- 2026-10-07
- Last updated
- 2026-10-07
Files and Weights
3 files, 19.3 MB in total. The weights are 1 file totalling 19.3 MB in pth.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| brats20_3d_unet_best.pth | Weights | 19.3 MB | 3e9fbc8be15b |
| README.md | Documentation | 1.7 KB | — |
| .gitattributes | Repository | 1.5 KB | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 19.3 MB
Released by Sagar Kumar through its official repository on Hugging Face.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 19.3 MB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.