AnalysisAI ModelsOctober 18, 2026

Community GGUF quant brings DeepSeek V4.1 Flash 1017 vision to ds4

Read original source →huggingface.co

A community contributor published a 2-bit GGUF quantization of DeepSeek V4.1 Flash 1017 for the ds4 runtime, matching the recipe used for the original DS4.1 Q2 build. The language model already loaded in ds4; the vision tower required a separate PR to work.

1 source

More stories today

Open the live feed