metadata
license: openrail++
This repository contains offset versions of https://huggingface.co/mhdang/dpo-sdxl-text2image-v1 and https://huggingface.co/mhdang/dpo-sd1.5-text2image-v1.
These can be added directly to any initialized UNet to inject DPO training into it. See the code below for usage (diffusers only.)
def inject_dpo(unet: UNet2DConditionModel, dpo_path: str, strict: bool = False) -> None:
"""
Injects DPO weights directly into your UNet.
Args:
unet (`UNet2DConditionModel`)
The initialized UNet from your pipeline.
dpo_path (`str`)
The path to the `.safetensors` file downloaded from https://huggingface.co/benjamin-paine/sd-dpo-offsets/.
Make sure you're using the right file for the right base model.
strict (`bool`, *optional*)
Whether or not to raise errors when a weight cannot be applied. Defaults to false.
"""
from safetensors import safe_open
with safe_open(dpo_offset_path, framework="pt", device="cpu") as f:
for key in f.keys():
key_parts = key.split(".")
current_layer = unet
for key_part in key_parts[:-1]:
current_layer = getattr(current_layer, key_part, None)
if current_layer is None:
break
if current_layer is None:
if strict:
raise IOError(f"Couldn't find a layer to inject key {key} in.")
continue
layer_param = getattr(current_layer, key_parts[-1], None)
if layer_param is None:
if strict:
raise IOError(f"Couldn't get weighht parameter for key {key}")
layer_param.data += f.get_tensor(key)
Now you can use this function like so:
from diffusers import StableDiffusionPipeline
import huggingface_hub
import torch
# load sdv15 pipeline
model_id = "Lykon/dreamshaper-8"
pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)
# download DPO offsets
dpo_path = huggingface_hub.hf_hub_download("painebenjamin/sd-dpo-offsets", "sd_v15_unet_dpo_offset.safetensors")
# inject
inject_dpo(pipe.unet, dpo_path)
# make image
prompt = "Two cats playing chess on a tree branch"
image = pipe(prompt, guidance_scale=7.5).images[0]
image.save("cats_playing_chess.png")
Or for XL:
from diffusers import StableDiffusionXLPipeline
# load sdxl pipeline
model_id = "Lykon/dreamshaper-xl-1-0"
pipe = StableDiffusionXLPipeline.from_pretrained(model_id, torch_dtype=torch.float16, variant="fp16")
# download DPO offsets
dpo_path = huggingface_hub.hf_hub_download("painebenjamin/sd-dpo-offsets", "sd_xl_unet_dpo_offset.safetensors")
# inject
inject_dpo(pipe.unet, dpo_path)
# make image
prompt = "Two cats playing chess on a tree branch"
image = pipe(prompt, guidance_scale=7.5).images[0]
image.save("cats_playing_chess.png")