SuperLogic

SuperLogic's activity

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

New activity in Nanbeige/Nanbeige4.3-33B 1 day ago

It's good, but too slow. Next step, quantize it into three values and put it into 8GB of VRAM. For now, everyone can use Qwen3.8-27B-Flash-MTP-GGUF.

1

#17 opened 1 day ago by

SuperLogic

New activity in SuperLogic/Qwen3.8-27B-Flash-MTP-GGUF 3 days ago

88K tokens on an 8 GB card - the K/V cache finally fits in VRAM

#5 opened 3 days ago by

SuperLogic

Does this quant model handle agent tool calls well?

1

#4 opened 4 days ago by

ghostplant

New activity in SuperLogic/Qwen3.8-27B-Flash-MTP-GGUF 9 days ago

吓人，才刚放上去就有1000多下载，下一步，准备把推理速度优化到下载量，先定个小目标1000t/s

#2 opened 9 days ago by

SuperLogic

New activity in prism-ml/Ternary-Bonsai-2-27B-gguf 11 days ago

Crazy fast, already optimized to over 70 tokens/s, can be pushed up to 300 tokens/s, on an 8GB 4060.狂飙，已经被优化到70token/s以上,修改最高300token/s，在8G的4060

2

#96 opened 12 days ago by

SuperLogic

New activity in prism-ml/Ternary-Bonsai-2-27B-gguf 12 days ago

在4060笔记本，8G下，优化到45Token/S，已经达到大厂的Token接口速度，彻底实现Token自由，大家再也不需要去大厂订阅。

7

#39 opened about 1 month ago by

SuperLogic

在4060笔记本，8G下，优化到45Token/S之后，再提10%到50Token/S，但这还不是上限...理论上限可能高达65Token/S

👍 2

2

#46 opened about 1 month ago by

SuperLogic

One 27B model, two jobs — chat:50token/s + JEV decision engine:0.5s — on an 8 GB laptop

🧠 3

1

#65 opened about 1 month ago by

SuperLogic

New activity in nerkyor/Qwen3.8-27B-Coder390-EfficientThink-Opus5.5-GPT6Astra-Grok4.7-DSV4Pro-K3-SFT-RLOO-MTP-DFlash2 12 days ago

Qwen3.8-27B-Flash-MTP-GGUF--5.79 GB. Code ~70-90 t/s, editing ~100-200 t/s.

#12 opened 12 days ago by

SuperLogic

New activity in ConwayResearch/Underdog-Saluki-27B-1.0 12 days ago

Qwen3.8-27B-Flash-MTP-GGUF--5.79 GB. Code ~70-90 t/s, editing ~100-200 t/s.

#8 opened 12 days ago by

SuperLogic

New activity in Qwen/Qwen3.8-27B 12 days ago

Qwen3.8-27B-Flash-MTP-GGUF--5.79 GB. Code ~70-90 t/s, editing ~100-200 t/s.

#213 opened 12 days ago by

SuperLogic

New activity in SuperLogic/Qwen3.8-27B-Flash-MTP-GGUF 12 days ago

So scary-70token/s-200token/s

#1 opened 12 days ago by

SuperLogic

New activity in prism-ml/Ternary-Bonsai-2-27B-gguf 12 days ago

Sitting frozen like a nuclear bomb, the 8G laptop speeds up from 50 tokens straight to 70 tokens，8G再次从50Token/S大幅升到恐怖的70Token/s

5

#75 opened 30 days ago by

SuperLogic

New activity in prism-ml/Ternary-Bonsai-2-27B-gguf 29 days ago

Sitting frozen like a nuclear bomb, the 8G laptop speeds up from 50 tokens straight to 70 tokens，8G再次从50Token/S大幅升到恐怖的70Token/s

5

#75 opened 30 days ago by

SuperLogic

New activity in prism-ml/Ternary-Bonsai-2-27B-gguf about 1 month ago

One 27B model, two jobs — chat:50token/s + JEV decision engine:0.5s — on an 8 GB laptop

🧠 3

1

#65 opened about 1 month ago by

SuperLogic

在4060笔记本，8G下，优化到45Token/S之后，再提10%到50Token/S，但这还不是上限...理论上限可能高达65Token/S

👍 2

2

#46 opened about 1 month ago by

SuperLogic

在4060笔记本，8G下，优化到45Token/S，已经达到大厂的Token接口速度，彻底实现Token自由，大家再也不需要去大厂订阅。

7

#39 opened about 1 month ago by

SuperLogic

在4060笔记本，8G下，优化到45Token/S，已经达到大厂的Token接口速度，彻底实现Token自由，大家再也不需要去大厂订阅。

7

#39 opened about 1 month ago by

SuperLogic