Fame PyTorch ROCm Kit 🔥

Built with Courage. Built with Fire. Built with FAME.

🚀 About

This project provides a complete and clean build kit for compiling a working PyTorch ROCm version on AMD GPUs – without the usual headaches involving ROCm, HIP, OpenMP, or Git submodules.

You will find everything needed to build, install, and test PyTorch from source, generating a valid .whl package for your local machine.

📋 Project Structure

File/Folder	Purpose
`build_torch.sh`	Script to build PyTorch from source
`install_torch.sh`	Script to install the generated Wheel file
`test_torch.sh`	Script to test PyTorch import and ROCm status
`fame_torch_freeze.txt`	Environment freeze (package versions)
`pytorch-wheel/`	(Optional) Folder containing the `.whl` file
`MY_WHEELI_NOTES.md`	Notes related to the Wheel build
`PyTorch_ROCm_Build_Notes.md`	Extended ROCm build notes
`README.md`	This document – the Fame Masterplan

🛠️ Build Instructions

Activate your ROCm environment

source ~/rocm_env/bin/activate

Navigate to the project folder

cd fame-pytorch-kit/

Build PyTorch

bash build_torch.sh

The generated .whl file will be located in pytorch/dist/.

📦 Install Instructions

Quick Install

bash install_torch.sh

Manual Install

cd pytorch/dist/
pip install torch-2.8.0a0+gitc402b3b-cp312-cp312-linux_x86_64.whl

Make sure your Python environment (venv) is activated!

✅ Test Instructions

bash test_torch.sh

Or manually:

python -c "import torch; print(torch.cuda.is_available())"
python -c "import torch; print(torch.cuda.get_device_name(0))"
python -c "import torch; print(torch.version.hip)"

Expected: True, GPU Name, HIP version string.

📜 License

This project is licensed under the MIT License.

🤝 Credits

Fame Kit proudly built by sbeierle with 🛡️ Courage, 🔥 Fire, and 🚀 Fame.

Built with Fire. Built with Fame. 🛡️🔥

=======

fame-pytorch-kit

cbd4317 (first commit)