123 lines
4.8 KiB
Python
123 lines
4.8 KiB
Python
import yaml
|
|
import torch
|
|
from diffusers import StableDiffusionXLPipeline, DiffusionPipeline
|
|
from .pipeline import PhotoMakerStableDiffusionXLPipeline
|
|
import os
|
|
import sys
|
|
import diffusers
|
|
dif_version = str(diffusers.__version__)
|
|
dif_version_int= int(dif_version.split(".")[1])
|
|
dir_path = os.path.dirname(os.path.abspath(__file__))
|
|
path_dir = os.path.dirname(dir_path)
|
|
file_path = os.path.dirname(path_dir)
|
|
|
|
def get_instance_path(path):
|
|
instance_path = os.path.normpath(path)
|
|
if sys.platform == 'win32':
|
|
instance_path = path.replace('\\', "/")
|
|
return instance_path
|
|
|
|
yaml_path = get_instance_path(os.path.join(path_dir ,'config','models.yaml'))
|
|
original_config_file=get_instance_path(os.path.join(path_dir,'config','sd_xl_base.yaml'))
|
|
loras_path = get_instance_path(os.path.join(path_dir,"config","lora.yaml"))
|
|
|
|
def get_lora_dict():
|
|
# 打开并读取YAML文件
|
|
with open(loras_path, 'r', encoding="UTF-8") as stream:
|
|
try:
|
|
# 解析YAML文件内容
|
|
data = yaml.safe_load(stream)
|
|
|
|
# 此时 'data' 是一个Python字典，里面包含了YAML文件的所有数据
|
|
# print(data)
|
|
return data
|
|
|
|
except yaml.YAMLError as exc:
|
|
# 如果在解析过程中发生了错误，打印异常信息
|
|
print(exc)
|
|
|
|
datas = get_lora_dict()
|
|
lora_lightning_list = datas["lightning_xl_lora"]
|
|
#print(lora_lightning_list)
|
|
def get_models_dict():
|
|
# 打开并读取YAML文件
|
|
with open(yaml_path, 'r', encoding="UTF-8") as stream:
|
|
try:
|
|
# 解析YAML文件内容
|
|
data = yaml.safe_load(stream)
|
|
|
|
# 此时 'data' 是一个Python字典，里面包含了YAML文件的所有数据
|
|
# print(data)
|
|
return data
|
|
|
|
except yaml.YAMLError as exc:
|
|
# 如果在解析过程中发生了错误，打印异常信息
|
|
print(exc)
|
|
|
|
def load_models(model_info,_sd_type,device,photomaker_path,lora,lora_path,trigger_words,lora_scale):
|
|
path = model_info["path"]
|
|
single_files = model_info["single_files"]
|
|
use_safetensors = model_info["use_safetensors"]
|
|
model_type = model_info["model_type"]
|
|
|
|
if model_type == "txt2img":
|
|
if single_files:
|
|
if dif_version_int>=28:
|
|
pipe = StableDiffusionXLPipeline.from_single_file(
|
|
pretrained_model_link_or_path=path, original_config=original_config_file, torch_dtype=torch.float16)
|
|
|
|
else:
|
|
pipe = StableDiffusionXLPipeline.from_single_file(
|
|
pretrained_model_link_or_path=path, original_config_file=original_config_file, torch_dtype=torch.float16
|
|
)
|
|
|
|
else:
|
|
pipe = StableDiffusionXLPipeline.from_pretrained(
|
|
path, torch_dtype=torch.float16,use_safetensors=use_safetensors
|
|
)
|
|
|
|
pipe = pipe.to(device)
|
|
if lora != "none":
|
|
if lora in lora_lightning_list:
|
|
pipe.load_lora_weights(lora_path)
|
|
pipe.fuse_lora()
|
|
else:
|
|
pipe.load_lora_weights(lora_path, adapter_name=trigger_words)
|
|
pipe.fuse_lora(adapter_names=[trigger_words, ], lora_scale=lora_scale)
|
|
|
|
elif model_type == "img2img":
|
|
if single_files:
|
|
#print("loading from a single_files")
|
|
if dif_version_int>=28:
|
|
pipe = PhotoMakerStableDiffusionXLPipeline.from_single_file(
|
|
pretrained_model_link_or_path=path, original_config=original_config_file, torch_dtype=torch.float16,use_safetensors=use_safetensors)
|
|
else:
|
|
pipe = PhotoMakerStableDiffusionXLPipeline.from_single_file(
|
|
pretrained_model_link_or_path=path, original_config_file=original_config_file, torch_dtype=torch.float16,use_safetensors=use_safetensors
|
|
)
|
|
else:
|
|
pipe = PhotoMakerStableDiffusionXLPipeline.from_pretrained(
|
|
path, torch_dtype=torch.float16,use_safetensors=use_safetensors
|
|
)
|
|
|
|
pipe = pipe.to(device)
|
|
pipe.load_photomaker_adapter(
|
|
os.path.dirname(photomaker_path),
|
|
subfolder="",
|
|
weight_name=os.path.basename(photomaker_path),
|
|
trigger_word="img" # define the trigger word
|
|
)
|
|
if lora != "none":
|
|
if lora in lora_lightning_list:
|
|
pipe.load_lora_weights(lora_path)
|
|
pipe.fuse_lora()
|
|
else:
|
|
pipe.load_lora_weights(lora_path, adapter_name=trigger_words)
|
|
pipe.fuse_lora(adapter_names=[trigger_words, ], lora_scale=lora_scale)
|
|
|
|
|
|
else:
|
|
raise f"using{model_type}node,must choice{model_type}type in model_loader node"
|
|
return pipe
|
|
|