I’ve been working on a TensorFlow Port of SD 1.5/2.x for the past year and definitely would like to be prepared for SDXL.

Does anyone know what the configuration will be for the new model/weights?