GitHub Explorer

by Alexey Ratnikov

GitHub Explorer

GitHub Explorer|TRENDING COMPARE|FEEDBACK

Back to search

Copyright (c) 2026 Alexey Ratnikov

kentstone84/PyTorch-2.10.0a0 - GitHub Explorer | GitHub Explorer | Trending | Compare

PyTorch-2.10.0a0

kentstone84•PUBLIC

Unlock full RTX 5080 performance in PyTorch! PyTorch does not support RTX 5080 (sm_120) natively, so I built custom CUDA 12.8 drivers and PyTorch binaries to make it work. This repo contains build scripts, benchmarks, and installation guides for running AI models at max efficiency on the RTX 5080.

BSD 3-Clause "New" or "Revised" License

Created on Mar 20, 2025

Updated on Apr 3, 2026

Stars

59

Forks

6

Watchers

59

Open Issues

2

Repository Health Score

💛

70/100

Good

Overall repository health assessment

Score Breakdown

Activity

Active development - updated this week

30/30

100%

Community

59 stars, 6 forks

0/30

0%

Documentation

Has description, wiki, license

20/20

100%

Maintenance

3.4% issue ratio

20/20

100%

Health score is calculated based on activity, community engagement, documentation quality, and maintenance practices

Languages

Python

84.0%

Jupyter Notebook

9.0%

PowerShell

5.0%

Dockerfile

1.7%

Shell

0.3%

Issues Analytics

8

Total Issues

All time

2

Open

25% of total

6

Closed

75% of total

4mo

Avg Close Time

Avg response

Issues Activity: Last 6 months

Top Labels

Hottest Issues

1

#4 Unsure where to put the files

23

closed

2

#24 patch_config.json is a placeholder?

6

open

3

#3 What about RTX5070?

5

closed

4

#5 this is uneeded nightly has supported 5000 series for along time

1

1

closed

5

#2 error when executing patch_blackwell.sh

1

1

closed

Dependencies

No package.json found

This might not be a Node.js project

Top Contributors

1

kentstone84

User

34

commits

2

claude

User

22

commits

Recent Commits

Add update section and training loss details

Kent Stone•3 months ago

55423ecView on GitHub

Create CODE_OF_CONDUCT.md

Kent Stone•3 months ago

6a25ddaView on GitHub

Merge pull request #22 from kentstone84/claude/pytorch-cuda-compatibility-01PfmWHufyzmeUHUv1c45V4H

Kent Stone•4 months ago

9dcbcbfView on GitHub

Add comprehensive evidence upload and community verification framework

Claude•4 months ago

104e13dView on GitHub

Add evidence framework: Timeline and screenshot organization structure

Claude•4 months ago

54c3f11View on GitHub

Merge pull request #21 from kentstone84/claude/pytorch-cuda-compatibility-01PfmWHufyzmeUHUv1c45V4H

Kent Stone•4 months ago

a2ccf99View on GitHub

Clean up: Move PROPOSED_README_UPDATE.md to docs, fix all links

Claude•4 months ago

1f58176View on GitHub

Organize documentation: Move all .md files to docs/ directory

Claude•4 months ago

fb11975View on GitHub

Merge pull request #20 from kentstone84/claude/pytorch-cuda-compatibility-01PfmWHufyzmeUHUv1c45V4H

Kent Stone•4 months ago

3ef4864View on GitHub

Add automated driver patching system - ready for Ghidra data

Claude•4 months ago

c0d6f50View on GitHub

Add FUN_005e0020: The function that actually handles sm_120 strings

Claude•4 months ago

c218685View on GitHub

Clarify function hierarchy: FUN_00f682d0 calls patched subfunctions

Claude•4 months ago

50f3550View on GitHub

Add Ghidra screenshot evidence and complete analysis documentation

Claude•4 months ago

a150a0cView on GitHub

Merge pull request #19 from kentstone84/claude/pytorch-cuda-compatibility-01PfmWHufyzmeUHUv1c45V4H

Kent Stone•4 months ago

16c0fadView on GitHub

Add verified benchmark results: 60.8% performance improvement

Claude•4 months ago

fd05b64View on GitHub

View all commits