Model Format: BF16 (Bfloat16)

• Why use this version? The official version on Tensor.art uses the FP8 format to save space. However, FP8 loses precision during training. I uploaded this BF16 version specifically for creators who want the best possible quality when training their own LoRAs.

Description from the official Krea AI page on Hugging Face:

"Krea 2 Raw: This is the Krea 2 Raw checkpoint, its not recommended for inference use. Its a good base for finetuning or post-training for your own needs and domains. For example, one use-case is to train LoRAs on midtrain and directly use them on Krea 2 Turbo."