mflux-community/qwen-image-2-1-mflux-bf16

Same weights, repacked vs Qwen/Qwen-Image-2.1. Same weights in different files (re-saved, re-sharded or new metadata), plus 8 new tensors.

Tensors changed: other 0/4, Embeddings 0/2, Norms 0/66, Attention 0/128, MLP 0/96

On the Hub · Downloads on Model Pulse