The collection of consistency trajectory LoRAs was tested to work with NoobAI preferably, but it can also work with any other models.
hustle-distributionmatching.safetensors - base lora, expected CFG 7-10, most of regular and irregular samplers are working
hustle-softsuperpose.safetensors - soften base lora
hustle-vectorized.safetensors - as DMD2 lora is not working in a good way there is a DMD2 with capabilities to work with NoobAI + expected CFG 1-3 (more is not recommended cause matrix mismatch)
