Hello! I found an issue with the v1.0 LoRA file. Although the model page describes it as a Krea 2 LoRA, the downloaded .safetensors file appears to contain an SDXL LoRA structure (CLIP-L + CLIP-G + SDXL UNet, rank 16), rather than Krea 2 layers. In Forge Neo, it produces LoRA mismatch for KModel and has no effect on the generated image, even with both Krea 2 INT4 and INT8 ConvRot checkpoints. Could you please check whether the uploaded LoRA file is the correct one for Krea 2? Thank you!