User shares test images from Ideogram4 and Krea2 models

Reddit user Ant_6431 posted sample images generated with Ideogram4 and Krea2 models, using int8 quantized safetensors with convrot and Qwen3VL text encoders. The post showcases outputs from both models for comparison.