wwdsada/qwenimage21gguf
Same weights, converted vs Qwen/Qwen-Image-2.1. Same weights stored in another dtype (BF16 → BF16/F32), plus 32 new tensors.
Tensors changed: other 0/5, Embeddings 0/2, Norms 0/66, Attention 0/128, MLP 0/32
Same weights, converted vs Qwen/Qwen-Image-2.1. Same weights stored in another dtype (BF16 → BF16/F32), plus 32 new tensors.
Tensors changed: other 0/5, Embeddings 0/2, Norms 0/66, Attention 0/128, MLP 0/32