-
Notifications
You must be signed in to change notification settings - Fork 9
/
Copy pathutils.py
228 lines (205 loc) · 9.01 KB
/
utils.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
from contextlib import nullcontext
from pipeline import StableDiffusionPipelineAllinOne
from ppdiffusers import DPMSolverMultistepScheduler
from PIL import Image
import cv2
import numpy as np
from modules.helper import lora_helper
import paddle
import os
import zipfile
# 基础模型，需要是paddle版本的权重，未来会加更多的权重
pretrained_model_name_or_path = "/home/aistudio/PPdiffusers-webui/models/NovelAI_latest_ab21ba3c_paddle"
# 我们加载safetensor版本的权重
lora_outputs_path = "text_encoder_unet_lora.safetensors"
# 加载之前的模型
pipe = StableDiffusionPipelineAllinOne.from_pretrained(pretrained_model_name_or_path, safety_checker=None, feature_extractor=None,
requires_safety_checker=False)
# 设置采样器，采样器移到这里实现
pipe.scheduler = DPMSolverMultistepScheduler.from_config(pipe.scheduler.config)
# 加载lora权重, 可以选择加载和不加载lora, 没有lora时注释下行
# pipe.apply_lora(lora_outputs_path)
# pipe.apply_lora()
support_scheduler = [
"EulerAncestralDiscrete",
"PNDM",
"DDIM",
"LMSDiscrete",
"HeunDiscrete",
"KDPM2AncestralDiscrete",
"KDPM2Discrete"
]
model_name_list = [
"Baitian/momocha",
"Linaqruf/anything-v3.0",
"MoososCap/NOVEL-MODEL",
"Baitian/momoco",
"hequanshaguo/monoko-e",
"ruisi/anything",
"hakurei/waifu-diffusion-v1-3",
"CompVis/stable-diffusion-v1-4",
"runwayml/stable-diffusion-v1-5",
"stabilityai/stable-diffusion-2",
"stabilityai/stable-diffusion-2-base",
"hakurei/waifu-diffusion",
"naclbit/trinart_stable_diffusion_v2_60k",
"naclbit/trinart_stable_diffusion_v2_95k",
"naclbit/trinart_stable_diffusion_v2_115k",
"ringhyacinth/nail-set-diffuser",
"Deltaadams/Hentai-Diffusion",
"BAAI/AltDiffusion",
"BAAI/AltDiffusion-m9",
"IDEA-CCNL/Taiyi-Stable-Diffusion-1B-Chinese-v0.1",
"IDEA-CCNL/Taiyi-Stable-Diffusion-1B-Chinese-EN-v0.1",
"huawei-noah/Wukong-Huahua"]
def get_size(standard_size):
if standard_size == '512x768':
width, height = 512, 768
elif standard_size == '768x512':
width, height = 768, 512
elif standard_size == '512x512':
width, height = 512, 512
elif standard_size == '640x640':
width, height = 640, 640
elif standard_size == '自动判断':
width, height = -1, -1
else:
width, height = 512, 512
return width, height
context_null = nullcontext()
def ReadImage(image, height=None, width=None):
"""
Read an image and resize it to (height,width) if given.
If (height,width) = (-1,-1), resize it so that
it has w,h being multiples of 64 and in medium size.
"""
if isinstance(image, str):
image = Image.open(image).convert('RGB')
# clever auto inference of image size
w, h = image.size
if height == -1 or width == -1:
if w > h:
width = 768
height = max(64, round(width / w * h / 64) * 64)
else: # w < h
height = 768
width = max(64, round(height / h * w / 64) * 64)
if width > 576 and height > 576:
width = 576
height = 576
if (height is not None) and (width is not None) and (w != width or h != height):
image = image.resize((width, height), Image.ANTIALIAS)
return image
def txt2img(pipe, prompt, scheduler_name, width, height, guidance_scale, num_inference_steps, negative_prompt,
max_embeddings_multiples, enable_parsing, fp16=False, seed=None):
# scheduler = pipe.create_scheduler(scheduler_name)
# 边运行的时候会边传递值到这里！
def callback_fn(i, total, tqdm_progess):
print(i, total)
# print(i, total, tqdm_progess.format_dict)
if fp16 and scheduler_name != "LMSDiscrete":
context = paddle.amp.auto_cast(True, level='O2')
else:
context = context_null
with context:
return pipe.text2image(
prompt,
seed=seed,
width=width,
height=height,
guidance_scale=guidance_scale,
num_inference_steps=num_inference_steps,
negative_prompt=negative_prompt,
max_embeddings_multiples=int(max_embeddings_multiples),
skip_parsing=(not enable_parsing),
# scheduler=scheduler,
callback=callback_fn,
).images[0]
def img2img(pipe, image_path, prompt, scheduler_name, height, width, strength, num_inference_steps, guidance_scale,
negative_prompt, max_embeddings_multiples, enable_parsing, fp16=True, seed=None):
# scheduler = pipe.create_scheduler(scheduler_name)
init_image = ReadImage(image_path, height=height, width=width)
# 边运行的时候会边传递值到这里！
def callback_fn(i, total, tqdm_progess):
print(i, total)
# print(i, total, tqdm_progess.format_dict)
if fp16 and scheduler_name != "LMSDiscrete":
context = paddle.amp.auto_cast(True, level='O2')
else:
context = context_null
with context:
return pipe.img2img(prompt,
seed=seed,
image=init_image,
num_inference_steps=num_inference_steps,
strength=strength,
guidance_scale=guidance_scale,
negative_prompt=negative_prompt,
max_embeddings_multiples=int(max_embeddings_multiples),
skip_parsing=(not enable_parsing),
# scheduler=scheduler,
callback=callback_fn
).images[0]
def inpaint(pipe, image_path, mask_path, prompt, scheduler_name, height, width, num_inference_steps, strength,
guidance_scale, negative_prompt, max_embeddings_multiples, enable_parsing, fp16=True, seed=None):
# scheduler = pipe.create_scheduler(scheduler_name)
init_image = ReadImage(image_path, height=height, width=width)
mask_image = ReadImage(mask_path, height=height, width=width)
# 边运行的时候会边传递值到这里！
def callback_fn(i, total, tqdm_progess):
print(i, total)
# print(i, total, tqdm_progess.format_dict)
if fp16 and scheduler_name != "LMSDiscrete":
context = paddle.amp.auto_cast(True, level='O2')
else:
context = context_null
with context:
return pipe.inpaint(
prompt,
seed=seed,
image=init_image,
mask_image=mask_image,
num_inference_steps=num_inference_steps,
strength=strength,
guidance_scale=guidance_scale,
negative_prompt=negative_prompt,
max_embeddings_multiples=int(max_embeddings_multiples),
skip_parsing=(not enable_parsing),
# scheduler=scheduler,
callback=callback_fn
).images[0]
# train dreambooth lora, aistudio project: https://aistudio.baidu.com/aistudio/projectdetail/5481677
# TODO 在训练中可视化输出的图到UI界面，拖动滑块查看不同时间步产生的结果
def train_dreambooth_lora(zip_file, pretrained_model_name_or_path="Baitian/momocha", instance_data_dir="./Xinhai", output_dir="./dream_booth_lora_outputs", instance_prompt="Xinhai", resolution=512, train_batch_size=1,
gradient_accumulation_steps=1, checkpointing_steps=50, learning_rate=1e-4, report_to="visualdl", lr_scheduler="constant", lr_warmup_steps=0,
max_train_steps=100, lora_rank=128, validation_prompt="Xinhai", validation_epochs=25, validation_guidance_scale=5.0, use_lion=False, seed=0):
def unzip_file(zip_file):
os.makedirs("train_dreambooth_lora", exist_ok=True)
with zipfile.ZipFile(zip_file) as zip_ref:
zip_ref.extractall('./train_dreambooth_lora')
def process_zip_file(zip_file):
unzip_file(zip_file.name)
print("文件已解压并处理完成。")
# 在这里添加您的处理代码
os.system(f'python modules/train_dreambooth_lora.py \
--pretrained_model_name_or_path={pretrained_model_name_or_path} \
--instance_data_dir="./train_dreambooth_lora" \
--output_dir={output_dir} \
--instance_prompt={instance_prompt} \
--resolution={resolution} \
--train_batch_size={train_batch_size} \
--gradient_accumulation_steps={gradient_accumulation_steps} \
--checkpointing_steps={checkpointing_steps} \
--learning_rate={learning_rate} \
--report_to={report_to} \
--lr_scheduler={lr_scheduler} \
--lr_warmup_steps={lr_warmup_steps} \
--max_train_steps={max_train_steps} \
--lora_rank={lora_rank} \
--validation_prompt={validation_prompt} \
--validation_epochs={validation_epochs} \
--validation_guidance_scale={validation_guidance_scale} \
--use_lion {use_lion} \
--seed={seed}')
return "训练完成！"
return process_zip_file(zip_file)