👋 Open to Work

Sailor丶JS

sailorjs0804

orcarouter

·

AI & ML interests

SpeechLM & Harness

Recent Activity

new activity 3 days ago

orcarouter/Qwen3.8-Flash-Next-Uncensored-GGUF:Qwen3.8-Flash-Next iQ3_M GGUF: MTP achieves ~60% acceptance but is slower than non-speculative decoding on a single RTX 5090

new activity 3 days ago

orcarouter/Qwen3.8-Flash-Next-Uncensored-GGUF:Can you please update the MTP draft model to be compatible with current llama.cpp upstream?

new activity 3 days ago

orcarouter/Qwen3.8-Flash-Next-Uncensored-GGUF:IQ4_XS with the MTP head merged — may we publish the repack?

View all activity

Organizations

sailorjs0804 's datasets

None public yet