Do you think the new training method is effective?
a lora
formidable(ponyV1).safetensors
formidable_ponyV1_.safetensors