is that true that this DeepSeek-V4.1-Flash-1017 · GSQ-RCO Coder model can be run on 12GB VRAM / local?!
https://huggingface.co/ISTA-DASLab/DeepSeek-V4.1-Flash-1017-GSQ-RCO-Coder-GGUF
huggingface.coISTA-DASLab/DeepSeek-V4.1-Flash-1017-GSQ-RCO-Coder-GGUF · Hugging FaceImage-Text-to-Text model with 197B parameters. Download, explore files and run it on Hugging Face.
commit 81be107