π“Ÿ host1.bitchan.one

𓀐 @Q=x3=nwp

⏱ 1w seenβ›“ 0πŸ‘ 0β§– 1

filter [-]

languages:onother

How is /lmg/ using their new GPU cache commands in llmao.cpp with GLM 5.3 and Qwen Flash Next? What kind of results are you seeing? https://github.com/ggml-org/llama.cpp/pull/29887⏱ 1w↩ 1
10cm of GLM's height is hair.⏱ 1w↩