mst3c_z-image: v1.0: mst3c_z-image_000003750.safetensors

Based on the same dataset as my earlier Flux and SDXL models, works well at 0.8-1.0 strength when no mention of 'drawing' is in the prompt. Otherwise, if you add 'drawing', it will work better at 0.5-0.7 strength.