Running Qwen3.8 at 255 tok/s on a single 5090 (meshapplied.com)

Qwen3-Next (qwen.ai)

519 points by tosh 1 year ago | 201 comments

Qwen team blog post announcing Qwen3-Next, a model architecture aimed at improving LLM training and inference efficiency.

qwen qwen3-next llm mixture-of-experts

Screenshot of github.com: There is no official XScreenSaver GitHub repository

There is no official XScreenSaver GitHub repository (github.com)

63 points by st_goliath 6 days ago | 66 comments

A GitHub repository by jwz whose only content is a README stating there is no official XScreenSaver GitHub repository.

xscreensaver jwz github linux

Screenshot of unsloth.ai: How to run Qwen 3.5 locally

How to run Qwen 3.5 locally (unsloth.ai)

424 points by Curiositry 7 months ago | 138 comments

Unsloth documentation guide explaining how to run Alibaba's Qwen3.5 model family locally using GGUF quantized versions, with updates on improved quantization.

qwen unsloth local-llm gguf

Screenshot of github.com: Run Kimi K3 using 29 GB of RAM at 0.50 tok/s

Run Kimi K3 using 29 GB of RAM at 0.50 tok/s (github.com)

318 points by marcobambini 2 months ago | 152 comments

GitHub repository for waste, a dependency-free embeddable C inference engine that runs the 2.78-trillion-parameter Kimi K3 model by streaming weights from NVMe.

kimi-k3 llm-inference c nvme-streaming