gguf-org

/

z-image-gguf

like 36

Follow

gguf 137

just roughly tested q4; almost identical

just roughly tested q4; almost identical

i mean the different Qwen versions (thinking, instruct, vl, ect)

i mean the different Qwen versions (thinking, instruct, vl, ect)

almost identical. already stated

Thank you Calcuis, for making these resources available to GPU poor folks like us. otherwise its very, very slow going on CPU. And thanks is due to all other hardworking creators, developers as well.

Hi, there.

I was doing some tests, hoping I could use the same GGUF model as CLIP and as a regular vision model as well, but Qwen3-VL-4B-Instruct-GGUF added a lot of artifacts to the final image although it was listed as compatible in the README... Using qwen3_4b_f32-q8_0.gguf from this repo worked as intended. Did I miss something?

I would also like to thank @calcius a lot for his work and models, much appreciated.

Regards