Yohann Abouth

YoRandom

·

AI & ML interests

None yet

Recent Activity

new activity about 19 hours ago

meta-models/Muse-Glimmer-30B:Muse Glimmer 30B works pretty well with 512k context as-is

new activity 1 day ago

Qwen/Qwen3.8-27B:Benchmarked Qwen3.8-27B vs Nemotron 3.5 Lightning and Muse Glimmer on 16 hard problems (RTX 5090, Q4_K_M)

new activity 3 days ago

meta-models/Muse-Glimmer-30B:one week of Muse Glimmer in production, guided decoding at high volume (vLLM, single RTX 5090)

View all activity

Organizations

New activity in meta-models/Muse-Glimmer-30B about 19 hours ago

Muse Glimmer 30B works pretty well with 512k context as-is

#59 opened 3 days ago by

New activity in Qwen/Qwen3.8-27B 1 day ago

Benchmarked Qwen3.8-27B vs Nemotron 3.5 Lightning and Muse Glimmer on 16 hard problems (RTX 5090, Q4_K_M)

#103 opened 6 days ago by

New activity in meta-models/Muse-Glimmer-30B 3 days ago

one week of Muse Glimmer in production, guided decoding at high volume (vLLM, single RTX 5090)

#56 opened 7 days ago by

Multi-digit literals come back with spaces between the digits (`42` → `4 2`): generation, not tokenization. Anyone else seeing this?

#58 opened 3 days ago by

New activity in Qwen/Qwen3.8-27B 3 days ago

Horrible model .... Overthink eats all context tokens

#136 opened 4 days ago by

New activity in nvidia/Qwen3.6-35B-A3B-NVFP4 3 months ago

[Bug] Three stability failure modes on RTX 5090 (consumer Blackwell, SM120) — illegal instruction, CUDA-graph-replay crash, and silent engine hang

#9 opened 3 months ago by