Base Model : SD 1.5
If any issue full repo here
Trained for more than 24 hrs with 111 images of resolution 512.

You can use lora or textual inversion for better results and use sampler DPM++ 2M Karras.