Yohann Abouth

YoRandom's activity

• All
• Models
• Datasets
• Spaces
• Buckets
• Papers
• Collections
• Community
• Posts
• Upvotes
• Likes
• Articles

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

So, Mark? MUSE SPARK 1.2 ?

👍 1

1

#62 opened 21 days ago by

RomanHammer

New activity in Qwen/Qwen3.8-27B 19 days ago

Qwen3.8 isn't a general purpose AI model.

👀👍 10

9

#179 opened 24 days ago by

phil111

New activity in meta-models/Muse-Glimmer-30B 25 days ago

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

🚀👍 7

2

#56 opened about 1 month ago by

YoRandom

New activity in Qwen/Qwen3.8-27B 26 days ago

Horrible model .... Overthink eats all context tokens

👍😔 6

23

#136 opened about 1 month ago by

kremerneil

New activity in meta-models/Muse-Glimmer-30B about 1 month ago

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

2

#59 opened about 1 month ago by

lobanov

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

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

🔥❤️ 11

13

#103 opened about 1 month ago by

laxmimerit

New activity in meta-models/Muse-Glimmer-30B about 1 month ago

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

2

#58 opened about 1 month ago by

buley

New activity in nvidia/Qwen3.6-35B-A3B-NVFP4 4 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 4 months ago by

YoRandom