Closed bennyguo closed 1 year ago
The LoRA should always take conditional text embeddings as it is trained so. Not affecting the results when guidance_scale_lora=1.
guidance_scale_lora=1
The LoRA should always take conditional text embeddings as it is trained so. Not affecting the results when
guidance_scale_lora=1
.