pitill0

pitill0/legacy-gpu-llm

Give older GPUs a second life running local LLMs with reproducible containers, hardware profiles and low-VRAM configurations.

0 stars 0 forks 0 watchers Shell MIT License
containers cuda gpu-inference legacy-hardware llama-cpp local-llm low-vram nvidia self-hosted
View on GitHub
2 Open Issues Need Help Last updated: Jul 20, 2026

Open Issues Need Help

View All on GitHub
Looking for a collaborator with access to a real Slurm GPU cluster 5 days ago
help wanted
pitill0/legacy-gpu-llm
0

Give older GPUs a second life running local LLMs with reproducible containers, hardware profiles and low-VRAM configurations.

Shell
#containers#cuda#gpu-inference#legacy-hardware#llama-cpp#local-llm#low-vram#nvidia#self-hosted
Help us validate Vulkan on one more Linux GPU 2 months ago
help wanted hardware-validation
pitill0/legacy-gpu-llm
0

Give older GPUs a second life running local LLMs with reproducible containers, hardware profiles and low-VRAM configurations.

Shell
#containers#cuda#gpu-inference#legacy-hardware#llama-cpp#local-llm#low-vram#nvidia#self-hosted