Joe lafrite

joelafrite's activity

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

New activity in gratex/Qwen3.8-27B-DFlash2-W4A16-g128-sym-GPTQ about 4 hours ago

apply_dense_kv_fix.py: silent corruption risk with NVFP4 drafters (+ working variant)

3

#1 opened 1 day ago by

joelafrite

updated a model about 4 hours ago

joelafrite/Qwen3.8-27B-fp8-KV-scale-calibration

Updated about 4 hours ago

published a model about 4 hours ago

joelafrite/Qwen3.8-27B-fp8-KV-scale-calibration

Updated about 4 hours ago

liked 6 models about 21 hours ago

unsloth/GLM-5.3-Flash-GGUF

Text Generation • 321B • Updated 3 days ago • 63.7k • 314

unsloth/Qwen3.8-27B-GGUF

27B • Updated 12 days ago • 9.35M • 3.3k

unsloth/Qwen3.8-Flash-Next-GGUF

Image-Text-to-Text • 177B • Updated about 1 hour ago • 431k • 645

zai-org/GLM-5.3

Text Generation • 753B • Updated about 21 hours ago • 94.4k • • 1.43k

zai-org/GLM-5.3-Flash

Image-Text-to-Text • 321B • Updated about 23 hours ago • 441k • • 1.84k

Qwen/Qwen3.8-Flash-Next

Image-Text-to-Text • 180B • Updated 5 days ago • 208k • 4.57k

liked 2 models 1 day ago

LibertAIDAI/GLM-5.3-Flash-NVFP4

Image-Text-to-Text • 165B • Updated 2 days ago • 25k • 56

gratex/Qwen3.8-27B-DFlash2-W4A16-g128-sym-GPTQ

2B • Updated about 15 hours ago • 226 • 1

New activity in LibertAIDAI/GLM-5.3-Flash-NVFP4 1 day ago

Datapoint: fp4 vs fp8 GLM-5.3-Flash endpoint parity on GSM8K + HumanEval

#8 opened 1 day ago by

joelafrite

New activity in QUASAR-QAT/Qwen3.8-27B-QUASAR-NVFP4 3 days ago

Heads-up: config declares an MTP head but no mtp.* tensors ship (silent 0%-acceptance trap) — and a working graft recipe

2

#3 opened 3 days ago by

joelafrite

New activity in peculiar-ragdoll/Qwen-Sharp-Chat-Templates 3 days ago

Production adoption report (vLLM + MTP): terseness confirmed, quality up at n=4 — and the MTP think-leak did NOT reproduce

❤️ 1

1

#11 opened 3 days ago by

joelafrite

New activity in orcarouter/Qwen3.8-27B-Uncensored-NVFP4 3 days ago

vLLM boot failure: quantization target literal `lm_head` never matches — one-line fix

1

#3 opened 3 days ago by

joelafrite

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

v2 validated on vLLM + NVFP4 target (RTX 5090): AL ~3.22 confirms your claim — with a 32GB context warning

#8 opened 3 days ago by

joelafrite

New activity in maurienne-ai/Qwen3.8-27B-DFlash2-NVFP4-RTNcal 3 days ago

config.json `producer` string breaks vLLM draft-model loading (+ numbers on vLLM nightly)

1

#1 opened 3 days ago by

joelafrite

New activity in RadixArk/Qwen3.8-Flash-Next-NVFP4 5 days ago

Degeneration Loop when calling tool (PDF/Image)

1

#3 opened 5 days ago by

dangnsh

liked a model 12 days ago

orcarouter/Qwen3.8-27B-Uncensored-NVFP4

Image-Text-to-Text • 21B • Updated 5 days ago • 49.7k • 124

liked a model 17 days ago

Qwen/Qwen3.8-27B

Image-Text-to-Text • 28B • Updated 18 days ago • 4.96M • • 13.5k