tsqn/Z-Image-Turbo_fp32-fp16-bf16_full_and_ema-only
Same weights, converted vs Tongyi-MAI/Z-Image-Turbo. Same weights stored in another dtype (F32 → BF16).
Tensors changed: other 0/70, Embeddings 0/9, Norms 0/204, Attention 0/136, MLP 0/102
Same weights, converted vs Tongyi-MAI/Z-Image-Turbo. Same weights stored in another dtype (F32 → BF16).
Tensors changed: other 0/70, Embeddings 0/9, Norms 0/204, Attention 0/136, MLP 0/102