Skip to content
#

multi-gpu

Here are 150 public repositories matching this topic...

Qwen3.8-Flash-Next on 2× RTX 3090: with 128 GB RAM up to 4,191 tok/s prefill · 111.5 tok/s decode (131K prompt), full 256K window at 2,865 tok/s; with 64 GB RAM 3,410 tok/s prefill · 84 tok/s decode. New: opt-in uncensored mode (runtime abliteration, no new weights).

  • Updated Oct 15, 2026
  • Python

Efficient and Scalable Physics-Informed Deep Learning and Scientific Machine Learning on top of Tensorflow for multi-worker distributed computing

  • Updated Mar 1, 2022
  • Python
local-llm-tuning-guide

A decision library for running local LLMs on multi-GPU machines without NVLink: parallelism, communication, quantization, speculative decoding, KV cache. Every conclusion carries its evidence grade and reproducible conditions.

  • Updated Nov 12, 2026
  • Python
lilbee

A local model engine with a search engine and web crawler built in. Run your coding agents on your own GPUs, and get cited answers from your files, code, and the sites you crawl. One executable.

  • Updated Nov 22, 2026
  • Python

Add this topic to your repo

To associate your repository with the multi-gpu topic, visit your repo's landing page and select "manage topics."

Learn more