Get the latest tech news

Show HN: Maple – Sub-2-BPW Moe Pruning to Run Gemma 4 26B at 45 TPS on a 6GB GPU

MAPLE: Manifold-Activated Pruning of Latent Experts for Sub-2-BPW MoE LLMs - AmSach/maple

None

Get the Android app

Or read this on Hacker News

Related news:

Fine-tuned Qwen on my 16gb GPU to respond without the letter e

Gemma-4-26B-A4B with 39-69% less thinking at same accuracy

Behold, Intel plans a 480GB GPU, a video card with more memory than the average laptop SSD has storage — but of course it's for AI, and curiously, it will use LPPDR5X, not even GDDR7

« Kirlian Photography: Building Your Own Equipment

British soldier ignored order to pull his SIM card, then got blackmailed »