2026-10-23 · America/Los_Angeles · 社区动态 · #8

ISTA-DASLab/DeepSeek-V4.1-Flash-1017-GSQ-RCO

A brand-new release from ISTA-DASLab with their GSQ-RCO smart quantization approach. Now DeepSeek-V4.1-Flash-1017 The quants are available on Hugginface: https://huggingface.co/ISTA-DASLab/DeepSeek-V4.1-Flash-1017-GSQ-RCO-GGUF Llama.cpp is not supported because the original architecture support is not yet merged: https://github.com/ggml-org/llama.cpp/pull/28696 https://github.com/ggml-org/llama.cpp/pull/30817 Strata support is mentioned in the model card but I'm pretty sure the original Strata only supports Qwen3.8-Flash-Next so most likely what they meant was one of the forks based on…

热度 50.3 / 100；排名与评分保留该期记录。