deepseek-v4.1IMAGE
12GB 显卡本地跑 DeepSeek V4.1 Flash：RTX 5070 上 105–117 tok/s
用 GSQ-RCO 量化配 Strata 推理引擎，可在 12GB 消费级显卡上跑 DeepSeek-V4.1-Flash，RTX 5070 加 64GB 内存达 105–117 tok/s，prefill 明显快于 decode。
Md Ismail Šojal 🕷️
@0x0SojalSec5 天前收录原推发布于 2026.10.25
Md Ismail Šojal 🕷️ @0x0SojalSec原推发布于 2026.10.25
You can now run a DeepSeek-V4.1-Flash locally 12GB GPU and Offloads consumer hardware. - GSQ-RCO quantizations with Strata inference engine. - RTX 5070 with 64GB/ 105-117 tok/s - Prefill is much faster than decode not bad - More RAM to faster decode - engine uses layer-major chunked reading, - for fun is a great.
x.com/0x0sojalsec/status/2114179761874612255采集于 2026.10.25 11:59