mflux-community/qwen-image-mflux-bf16

Same weights, repacked vs Qwen/Qwen-Image-2512. Same weights in different files (re-saved, re-sharded or new metadata), plus 840 new tensors.

Tensors changed: other 0/6, Embeddings 0/4, Attention 0/840, Norms 0/242

On the Hub · Downloads on Model Pulse