natural-tiger-sdxl-dora-v0-3.safetensors here 💾.models/Lora folder.<lora:natural-tiger-sdxl-dora-v0-3:1> to your prompt. On ComfyUI just load it as a regular LoRA.natural-tiger-sdxl-dora-v0-3_emb.safetensors here 💾.embeddings foldernatural-tiger-sdxl-dora-v0-3_emb to your prompt. For example, a natural-tiger-sdxl-dora-v0-3_emb tiger
(you need both the LoRA and the embeddings as they were trained together for this LoRA)from diffusers import AutoPipelineForText2Image
import torch
from huggingface_hub import hf_hub_download
from safetensors.torch import load_file
pipeline = AutoPipelineForText2Image.from_pretrained('stabilityai/stable-diffusion-xl-base-1.0', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('thliang01/natural-tiger-sdxl-dora-v0-3', weight_name='pytorch_lora_weights.safetensors')
embedding_path = hf_hub_download(repo_id='thliang01/natural-tiger-sdxl-dora-v0-3', filename='natural-tiger-sdxl-dora-v0-3_emb.safetensors', repo_type="model")
state_dict = load_file(embedding_path)
pipeline.load_textual_inversion(state_dict["clip_l"], token=["<s0>", "<s1>"], text_encoder=pipeline.text_encoder, tokenizer=pipeline.tokenizer)
pipeline.load_textual_inversion(state_dict["clip_g"], token=["<s0>", "<s1>"], text_encoder=pipeline.text_encoder_2, tokenizer=pipeline.tokenizer_2)
image = pipeline('a <s0><s1> tiger').images[0]
For more details, including weighting, merging and fusing LoRAs, check the documentation on loading LoRAs in diffusers
To trigger image generation of trained concept(or concepts) replace each concept identifier in you prompt with the new inserted tokens:
to trigger concept SBU → use <s0><s1> in your prompt
All Files & versions.
The weights were trained using 🧨 diffusers Advanced Dreambooth Training Script.
LoRA for the text encoder was enabled. False.
Pivotal tuning was enabled: True.
Special VAE used for training: madebyollin/sdxl-vae-fp16-fix.
Base model
stabilityai/stable-diffusion-xl-base-1.0