torch-cu128-sm120 (RTX 5080) Offline

⚡ Prebuilt PyTorch + torchvision for Ada GPUs (e.g., RTX 5080)
🔧 free-VPN · Built for CUDA 12.8 · Verified in Gaussian Splatting / SplaTAM
🎓 Educational: How to compile PyTorch for sm_120.
😣 Feedback & Issues: Use GitHub Issues
🙌 Contribute: See CONTRIBUTING.md for guidelines

---

🚀 Quick Start （3 Steps)

---

📥 Step 1: Download .whl files

Package	Version (tag)	File Size	SHA256
torch	2.9.0a0+gitc665594	235 MB	d8534c88263d8f184afea91dfc4feaf8ee2038ce526861a9f8868e716f8a158
torchvision	0.24.0a0+f52c4f1	1.87 MB	40655e9c93425751e400de656f0c0ceb81383bfbe4f392edce5b56458892fb9f

💽 Step 2: environment

conda create -n torch_env python=3.10 -y
conda activate torch_env

pip install ./torch-2.9.0a0+gitc665594-cp310-cp310-linux_x86_64.whl
pip install ./torchvision-0.24.0a0+f52c4f1-cp310-linux_x86_64.whl

✅ Step 3: Verify

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

---

📌 Check

2.9.0a0+gitc665594
True
NVIDIA GeForce RTX 5080 Laptop GPU

---

📂 Project Structure

├── .github/                       # GitHub workflows & issue templates
│ ├── workflows/
│ │ └── test-install.yml           # CI workflow: test install from .whl
│ └── ISSUE_TEMPLATE/
│   ├── bug_report.yaml            # Issue template: bug report
│   ├── config.yml                 # Issue template config
│   └── feature_request.yaml       # Issue template: feature request
│
├── examples/
│ ├── demo_tensor_cuda.py          # Demo: test tensor on GPU
│ └── verify_install.py            # Verifies torch installation & CUDA
│
├── scripts/
│ ├── build_pytorch.sh              # Script to build PyTorch from source
│ └── upload_release.py             # Helper to upload .whl to GitHub release
│
├── requirements.txt                # Python dependency list (no torch)
├── install_from_whl.sh             # One-line installer using .whl
├── build_info.md                   # Build config & environment metadata
├── CHANGELOG.md                    # Version history
├── CONTRIBUTING.md                 # Contribution guidelines
├── LICENSE                         # License (MIT/Apache/etc.)
├── .gitignore                      # Git ignored files
└── README.md                       # Project overview & installation guide

📓 Changelog

• 👉 See CHANGELOG.md for update history.
• 👉 See build_info.md for full build details (OS, CUDA, commit hash, etc).