Krea2 quantization comparison: BF16, FP8, INT8, GGUF, MXFP8, NVFP4

Reddit user y3kdhmbdb2ch2fc6vpm2 generated 120 images across 20 sets comparing Krea2 turbo model quantizations (BF16, FP8, INT8, GGUF, MXFP8, NVFP4). The comparison helps users evaluate quality vs. performance trade-offs.