Kevin Gascoigne

Kjay

·

AI & ML interests

Graphic Design

Recent Activity

new activity 7 days ago

nvidia/Qwen3.6-27B-NVFP4:Works via Ollama, but the ~40 tok/s MTP speedup requires raw llama-server (Ollama has no --spec-type flag) - plus vLLM lockup findings on DGX Spark

new activity 29 days ago

joaquinpedreno/hello_world:Local LLM option?

liked a model about 1 month ago

nhathoangfoto/Flux.2-Klein-9B-SmartCharacterSwap

View all activity

Organizations

Kjay 's datasets

None public yet