-
Notifications
You must be signed in to change notification settings - Fork 10
/
app.py
274 lines (229 loc) · 11.8 KB
/
app.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
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
import os
from datetime import datetime as date_time
import sys
import gc
path = os.path.abspath("src")
sys.path.append(path)
import time
import torch
from diffusers import (
StableCascadeDecoderPipeline,
StableCascadePriorPipeline,
StableCascadeUNet,
)
from diffusers import LCMScheduler # LCM Scheduler
import gradio as gr
import random
from PIL import ImageEnhance
import image_save_file
from dotenv import load_dotenv
import platform
if torch.cuda.is_available():
device = "cuda"
elif torch.backends.mps.is_available():
device = "mps"
else:
device = "cpu"
torch_dtype = torch.bfloat16
def constrast_image(image_file, factor):
im_constrast = ImageEnhance.Contrast(image_file).enhance(factor)
return im_constrast
def generate_image(checkpoint_basename,checkpoint_prior,checkpoint_decoder,prompt_input,dynamic_prompt,negative_prompt,sampler_choice,num_images_per_prompt,random_seed,input_seed,width,height,guidance_scale,num_inference_steps,num_inference_steps_decoder,contrast):
def remove_last_comma(sentence):
if len(sentence) > 0 and sentence[-1] == ',':
sentence_without_comma = sentence[:-1]
return sentence_without_comma
else:
return sentence
def remove_duplicates(words):
words_list = words.split(",")
unique_words = []
for word in words_list:
if word not in unique_words:
unique_words.append(word)
unique_string = ",".join(unique_words)
return unique_string
if dynamic_prompt > 0:
if prompt_input != "":
if prompt_input[-1] != ",":
prompt_input = prompt_input + ","
banned_words = os.getenv("banned_words", "").split(",")
import app_retnet
prompt = app_retnet.main_def(prompt_input=prompt_input, max_tokens=dynamic_prompt, DEVICE="cpu", banned_words=banned_words, prompt_chara=False)
prompt = remove_duplicates(prompt.lower())
prompt = remove_last_comma(prompt)
else:
prompt = prompt_input
if prompt == "":
prompt = "a cat with the sign: prompt not found, write in black"
negative_prompt = negative_prompt
if random_seed:
input_seed = random.randint(0, 9999999999)
else:
input_seed = int(input_seed)
if float(guidance_scale).is_integer():
guidance_scale = int(guidance_scale) # for txt_file_data correct format
generator = torch.Generator(device=device).manual_seed(input_seed)
print(f"Prompt: {prompt}")
checkpoint_prior_name = ""
if len(checkpoint_prior) < 1:
prior = StableCascadePriorPipeline.from_pretrained("stabilityai/stable-cascade-prior", torch_dtype=torch_dtype).to(device)
checkpoint_prior_name = "stable_cascade"
else:
prior_unet = StableCascadeUNet.from_single_file(checkpoint_basename + checkpoint_prior,torch_dtype=torch_dtype)
prior = StableCascadePriorPipeline.from_pretrained("stabilityai/stable-cascade-prior", prior=prior_unet, torch_dtype=torch_dtype).to(device)
checkpoint_prior_name = os.path.splitext(checkpoint_prior)[0]
prior.safety_checker = None
prior.requires_safety_checker = False
resize_pixel_w = width % 128
resize_pixel_h = height % 128
if resize_pixel_w > 0:
width = width - resize_pixel_w
if resize_pixel_h > 0:
height = height - resize_pixel_h
start_time = time.time()
match sampler_choice:
case "LCM":
sampler = "LCM"
prior.scheduler = LCMScheduler.from_config(prior.scheduler.config)
case _:
sampler = "DDPMWuerstchenScheduler" # default
prior_output = prior(
prompt=prompt,
negative_prompt=negative_prompt,
generator=generator,
width=width,
height=height,
guidance_scale=guidance_scale,
num_inference_steps=num_inference_steps,
num_images_per_prompt=num_images_per_prompt
)
if len(checkpoint_prior) > 0:
del prior_unet
del prior
gc.collect()
if device=="cuda":
torch.cuda.empty_cache()
# checkpoint_decoder_name = ""
if len(checkpoint_decoder) < 1:
decoder = StableCascadeDecoderPipeline.from_pretrained("stabilityai/stable-cascade", torch_dtype=torch_dtype).to(device)
# checkpoint_decoder_name = "stable_cascade_decoder"
else:
decoder_unet = StableCascadeUNet.from_single_file(checkpoint_basename + checkpoint_decoder,torch_dtype=torch_dtype)
decoder = StableCascadeDecoderPipeline.from_pretrained("stabilityai/stable-cascade", decoder=decoder_unet, torch_dtype=torch_dtype).to(device)
# checkpoint_decoder_name = os.path.splitext(checkpoint_decoder)[0]
decoder.safety_checker = None
decoder.requires_safety_checker = False
images = decoder(image_embeddings=prior_output.image_embeddings,
prompt=prompt,
negative_prompt=negative_prompt,
generator=generator,
guidance_scale=0,
num_inference_steps=num_inference_steps_decoder,
output_type="pil"
).images
end_time = time.time()
duration = end_time - start_time
print(f"Time: {duration} seconds.")
if resize_pixel_w > 0:
width = width + resize_pixel_w
if resize_pixel_h > 0:
height = height + resize_pixel_h
for image in images:
if resize_pixel_w > 0 or resize_pixel_h > 0:
image = image.resize((width, height))
if contrast != 1:
image = constrast_image(image, contrast)
txt_file_data=prompt+"\n"+"Negative prompt: "+negative_prompt+"\n"+"Steps: "+str(num_inference_steps)+", Sampler: "+sampler+", CFG scale: "+str(guidance_scale)+", Seed: "+str(input_seed)+", Size: "+str(width)+"x"+str(height)+", Model: "+checkpoint_prior_name
file_path = image_save_file.save_file(image, txt_file_data)
if len(checkpoint_decoder) > 0:
del decoder_unet
del decoder
gc.collect()
if device=="cuda":
torch.cuda.empty_cache()
return_txt_file_data = f"{txt_file_data}\nTime: {duration} seconds."
yield images, return_txt_file_data
def stop_gen(checkpoint_prior, checkpoint_decoder):
try:
if len(checkpoint_prior) > 0:
del prior_unet
if len(checkpoint_decoder) > 0:
del decoder_unet
del prior
del decoder
gc.collect()
torch.cuda.empty_cache()
finally:
os.execv(sys.executable, [sys.executable, __file__, "restart"])
def open_dir(dir="image"):
current_datetime = date_time.now()
current_date = current_datetime.strftime(f"%Y_%m_%d")
folder = os.getcwd() + "/" + dir + "/" + current_date
if not os.path.exists(folder):
folder = os.getcwd() + "/" + dir
if os.path.exists(folder):
operating_system = platform.system()
if operating_system == 'Windows':
os.startfile(folder)
elif operating_system == 'Darwin':
os.system('open "{}"'.format(folder))
elif operating_system == 'Linux':
os.system('xdg-open "{}"'.format(folder))
if __name__ == "__main__":
load_dotenv("./env/.env")
default_checkpoint_basename=os.getenv("checkpoint_basename","./stable_cascade/")
default_negative_prompt = os.getenv("negative_prompt", "")
default_sampler = os.getenv("sampler", "DDPMWuerstchenScheduler")
default_batch_size = int(os.getenv("batch_size", "1"))
default_random_seed = os.getenv("random_seed", "true").lower() == "true"
default_input_seed = int(os.getenv("input_seed", "1234"))
default_width = int(os.getenv("width", "768"))
default_height = int(os.getenv("height", "1024"))
default_guidance_scale = float(os.getenv("guidance_scale", "4"))
default_num_inference_steps = int(os.getenv("num_inference_steps", "20"))
default_num_inference_steps_decoder = int(os.getenv("num_inference_steps_decode", "12"))
default_contrast = float(os.getenv("contrast", "1"))
sampler_choice_list= ["DDPMWuerstchenScheduler", "LCM"]
dynamic_prompt=int(os.getenv("dynamic_prompt", "0"))
generator_image = generate_image
inbrowser_ = True
if len(sys.argv) > 1:
if sys.argv[1] == "restart":
inbrowser_ = False
if default_checkpoint_basename[-1] != "/":
default_checkpoint_basename = default_checkpoint_basename + "/"
checkpoints_prior_list = [os.path.basename(file) for file in os.listdir(default_checkpoint_basename) if file.endswith(".safetensors")]
checkpoints_decoder_list = [os.path.basename(file) for file in os.listdir(default_checkpoint_basename) if file.endswith(".safetensors")]
with gr.Blocks() as demo:
with gr.Row():
with gr.Column():
title="stable_cascade_easy"
checkpoint_basename = gr.Textbox(value=default_checkpoint_basename, label="Checkpoint Path", visible=False)
checkpoint_prior=gr.Dropdown(value=None, choices=checkpoints_prior_list, allow_custom_value=True, filterable=True, label="Checkpoint(Prior, Stage C), empty for Stable Cascade Default Prior")
checkpoint_decoder=gr.Dropdown(value=None, choices=checkpoints_decoder_list, allow_custom_value=True, filterable=True, label="Checkpoint(Decoder, Stage B), empty for Stable Cascade Default Decoder")
prompt_input=gr.Textbox(value="", lines=4, label="Prompt")
dynamic_prompt = gr.Number(value=dynamic_prompt, label="Magic Prompt(max tokens, 0=off)",step=32,minimum=0,maximum=1024)
negative_prompt=gr.Textbox(value=default_negative_prompt, lines=4, label="Negative Prompt")
sampler_choice=gr.Dropdown(value=default_sampler, choices=sampler_choice_list, label="Scheduler")
num_images_per_prompt=gr.Number(value=default_batch_size, label="Batch Size",step=1,minimum=1,maximum=16)
random_seed=gr.Checkbox(value=default_random_seed, label="Random Seed")
input_seed=gr.Number(value=default_input_seed, label="Input Seed",step=1,minimum=0, maximum=9999999999)
width=gr.Number(value=default_width, label="Width",step=100)
height=gr.Number(value=default_height, label="Height",step=100)
guidance_scale=gr.Number(value=default_guidance_scale, label="Guidance Scale",step=1)
with gr.Row():
num_inference_steps=gr.Number(value=default_num_inference_steps, label="Steps Prior",step=1)
num_inference_steps_decoder=gr.Number(value=default_num_inference_steps_decoder, label="Steps Decoder",step=1)
contrast=gr.Slider(value=default_contrast, label="Contrast(Default Value = 1)",step=0.05,minimum=0.5,maximum=1.5)
with gr.Row():
btn_stop_gen = gr.Button(value="Stop")
btn_generate = gr.Button(value="Generate")
with gr.Column():
output_images=gr.Gallery(allow_preview=True, preview=True, label="Generated Images", show_label=True)
btn_open_dir = gr.Button(value="Open Image Directory")
output_text=gr.Textbox(label="Metadata")
btn_generate.click(generator_image, inputs=[checkpoint_basename, checkpoint_prior,checkpoint_decoder,prompt_input,dynamic_prompt,negative_prompt,sampler_choice,num_images_per_prompt,random_seed,input_seed,width,height,guidance_scale,num_inference_steps,num_inference_steps_decoder,contrast],outputs=[output_images,output_text])
btn_open_dir.click(open_dir, inputs=[], outputs=[])
btn_stop_gen.click(stop_gen, inputs=[checkpoint_prior,checkpoint_decoder], outputs=[])
demo.launch(inbrowser=inbrowser_)