BestWishYsh
commited on
Commit
•
2a90ae7
1
Parent(s):
b82463b
Simplify the code
Browse files
app.py
CHANGED
@@ -1,5 +1,6 @@
|
|
1 |
import os
|
2 |
import copy
|
|
|
3 |
import torch
|
4 |
import random
|
5 |
import gradio as gr
|
@@ -7,7 +8,7 @@ from glob import glob
|
|
7 |
from omegaconf import OmegaConf
|
8 |
from safetensors import safe_open
|
9 |
from diffusers import AutoencoderKL
|
10 |
-
from diffusers import
|
11 |
from diffusers.utils.import_utils import is_xformers_available
|
12 |
from transformers import CLIPTextModel, CLIPTokenizer
|
13 |
|
@@ -66,7 +67,6 @@ device = torch.device('cuda:0')
|
|
66 |
|
67 |
class MagicTimeController:
|
68 |
def __init__(self):
|
69 |
-
|
70 |
# config dirs
|
71 |
self.basedir = os.getcwd()
|
72 |
self.stable_diffusion_dir = os.path.join(self.basedir, "ckpts", "Base_Model")
|
@@ -93,18 +93,11 @@ class MagicTimeController:
|
|
93 |
self.unet = UNet3DConditionModel.from_pretrained_2d(pretrained_model_path, subfolder="unet", unet_additional_kwargs=OmegaConf.to_container(self.inference_config.unet_additional_kwargs)).to(device)
|
94 |
self.text_model = CLIPTextModel.from_pretrained("openai/clip-vit-large-patch14")
|
95 |
self.unet_model = UNet3DConditionModel.from_pretrained_2d(pretrained_model_path, subfolder="unet", unet_additional_kwargs=OmegaConf.to_container(self.inference_config.unet_additional_kwargs))
|
96 |
-
|
97 |
-
# self.tokenizer = tokenizer
|
98 |
-
# self.text_encoder = text_encoder
|
99 |
-
# self.vae = vae
|
100 |
-
# self.unet = unet
|
101 |
-
# self.text_model = text_model
|
102 |
|
103 |
self.update_motion_module(self.motion_module_list[0])
|
104 |
self.update_motion_module_2(self.motion_module_list[0])
|
105 |
self.update_dreambooth(self.dreambooth_list[0])
|
106 |
|
107 |
-
|
108 |
def refresh_motion_module(self):
|
109 |
motion_module_list = glob(os.path.join(self.motion_module_dir, "*.ckpt"))
|
110 |
self.motion_module_list = [os.path.basename(p) for p in motion_module_list]
|
@@ -113,7 +106,7 @@ class MagicTimeController:
|
|
113 |
dreambooth_list = glob(os.path.join(self.personalized_model_dir, "*.safetensors"))
|
114 |
self.dreambooth_list = [os.path.basename(p) for p in dreambooth_list]
|
115 |
|
116 |
-
def update_dreambooth(self, dreambooth_dropdown):
|
117 |
self.selected_dreambooth = dreambooth_dropdown
|
118 |
|
119 |
dreambooth_dropdown = os.path.join(self.personalized_model_dir, dreambooth_dropdown)
|
@@ -124,26 +117,18 @@ class MagicTimeController:
|
|
124 |
converted_vae_checkpoint = convert_ldm_vae_checkpoint(dreambooth_state_dict, self.vae.config)
|
125 |
self.vae.load_state_dict(converted_vae_checkpoint)
|
126 |
|
127 |
-
|
128 |
-
|
129 |
-
|
130 |
-
|
131 |
-
torch.cuda.empty_cache()
|
132 |
-
torch.cuda.empty_cache()
|
133 |
-
torch.cuda.empty_cache()
|
134 |
-
torch.cuda.empty_cache()
|
135 |
converted_unet_checkpoint = convert_ldm_unet_checkpoint(dreambooth_state_dict, self.unet_model.config)
|
136 |
self.unet = copy.deepcopy(self.unet_model)
|
137 |
self.unet.load_state_dict(converted_unet_checkpoint, strict=False)
|
138 |
|
139 |
-
|
140 |
-
|
141 |
-
|
142 |
-
|
143 |
-
torch.cuda.empty_cache()
|
144 |
-
torch.cuda.empty_cache()
|
145 |
-
torch.cuda.empty_cache()
|
146 |
-
torch.cuda.empty_cache()
|
147 |
text_model = copy.deepcopy(self.text_model)
|
148 |
self.text_encoder = convert_ldm_clip_text_model(text_model, dreambooth_state_dict)
|
149 |
|
@@ -182,17 +167,23 @@ class MagicTimeController:
|
|
182 |
height_slider,
|
183 |
seed_textbox,
|
184 |
):
|
|
|
|
|
|
|
185 |
if self.selected_motion_module != motion_module_dropdown: self.update_motion_module(motion_module_dropdown)
|
186 |
if self.selected_motion_module != motion_module_dropdown: self.update_motion_module_2(motion_module_dropdown)
|
187 |
if self.selected_dreambooth != dreambooth_dropdown: self.update_dreambooth(dreambooth_dropdown)
|
188 |
|
|
|
|
|
|
|
189 |
if is_xformers_available(): self.unet.enable_xformers_memory_efficient_attention()
|
190 |
|
191 |
pipeline = MagicTimePipeline(
|
192 |
vae=self.vae, text_encoder=self.text_encoder, tokenizer=self.tokenizer, unet=self.unet,
|
193 |
scheduler=DDIMScheduler(**OmegaConf.to_container(self.inference_config.noise_scheduler_kwargs))
|
194 |
-
).to(device)
|
195 |
-
|
196 |
if int(seed_textbox) > 0: seed = int(seed_textbox)
|
197 |
else: seed = random.randint(1, 1e16)
|
198 |
torch.manual_seed(int(seed))
|
@@ -225,16 +216,12 @@ class MagicTimeController:
|
|
225 |
"seed": seed,
|
226 |
"dreambooth": dreambooth_dropdown,
|
227 |
}
|
|
|
|
|
|
|
228 |
return gr.Video(value=save_sample_path), gr.Json(value=json_config)
|
229 |
|
230 |
-
|
231 |
-
# tokenizer = CLIPTokenizer.from_pretrained(pretrained_model_path, subfolder="tokenizer")
|
232 |
-
# text_encoder = CLIPTextModel.from_pretrained(pretrained_model_path, subfolder="text_encoder").cuda()
|
233 |
-
# vae = AutoencoderKL.from_pretrained(pretrained_model_path, subfolder="vae").cuda()
|
234 |
-
# unet = UNet3DConditionModel.from_pretrained_2d(pretrained_model_path, subfolder="unet", unet_additional_kwargs=OmegaConf.to_container(inference_config.unet_additional_kwargs)).cuda()
|
235 |
-
# text_model = CLIPTextModel.from_pretrained("openai/clip-vit-large-patch14")
|
236 |
-
# controller = MagicTimeController(tokenizer=tokenizer, text_encoder=text_encoder, vae=vae, unet=unet, text_model=text_model)
|
237 |
-
controller = MagicTimeController()
|
238 |
|
239 |
def ui():
|
240 |
with gr.Blocks(css=css) as demo:
|
@@ -255,9 +242,6 @@ def ui():
|
|
255 |
dreambooth_dropdown = gr.Dropdown( label="DreamBooth Model", choices=controller.dreambooth_list, value=controller.dreambooth_list[0], interactive=True )
|
256 |
motion_module_dropdown = gr.Dropdown( label="Motion Module", choices=controller.motion_module_list, value=controller.motion_module_list[0], interactive=True )
|
257 |
|
258 |
-
dreambooth_dropdown.change(fn=controller.update_dreambooth, inputs=[dreambooth_dropdown], outputs=[dreambooth_dropdown])
|
259 |
-
motion_module_dropdown.change(fn=controller.update_motion_module, inputs=[motion_module_dropdown], outputs=[motion_module_dropdown])
|
260 |
-
|
261 |
prompt_textbox = gr.Textbox( label="Prompt", lines=3 )
|
262 |
negative_prompt_textbox = gr.Textbox( label="Negative Prompt", lines=3, value="worst quality, low quality, nsfw, logo")
|
263 |
|
@@ -290,7 +274,6 @@ def ui():
|
|
290 |
|
291 |
return demo
|
292 |
|
293 |
-
|
294 |
if __name__ == "__main__":
|
295 |
demo = ui()
|
296 |
demo.queue(max_size=20)
|
|
|
1 |
import os
|
2 |
import copy
|
3 |
+
import time
|
4 |
import torch
|
5 |
import random
|
6 |
import gradio as gr
|
|
|
8 |
from omegaconf import OmegaConf
|
9 |
from safetensors import safe_open
|
10 |
from diffusers import AutoencoderKL
|
11 |
+
from diffusers import DDIMScheduler
|
12 |
from diffusers.utils.import_utils import is_xformers_available
|
13 |
from transformers import CLIPTextModel, CLIPTokenizer
|
14 |
|
|
|
67 |
|
68 |
class MagicTimeController:
|
69 |
def __init__(self):
|
|
|
70 |
# config dirs
|
71 |
self.basedir = os.getcwd()
|
72 |
self.stable_diffusion_dir = os.path.join(self.basedir, "ckpts", "Base_Model")
|
|
|
93 |
self.unet = UNet3DConditionModel.from_pretrained_2d(pretrained_model_path, subfolder="unet", unet_additional_kwargs=OmegaConf.to_container(self.inference_config.unet_additional_kwargs)).to(device)
|
94 |
self.text_model = CLIPTextModel.from_pretrained("openai/clip-vit-large-patch14")
|
95 |
self.unet_model = UNet3DConditionModel.from_pretrained_2d(pretrained_model_path, subfolder="unet", unet_additional_kwargs=OmegaConf.to_container(self.inference_config.unet_additional_kwargs))
|
|
|
|
|
|
|
|
|
|
|
|
|
96 |
|
97 |
self.update_motion_module(self.motion_module_list[0])
|
98 |
self.update_motion_module_2(self.motion_module_list[0])
|
99 |
self.update_dreambooth(self.dreambooth_list[0])
|
100 |
|
|
|
101 |
def refresh_motion_module(self):
|
102 |
motion_module_list = glob(os.path.join(self.motion_module_dir, "*.ckpt"))
|
103 |
self.motion_module_list = [os.path.basename(p) for p in motion_module_list]
|
|
|
106 |
dreambooth_list = glob(os.path.join(self.personalized_model_dir, "*.safetensors"))
|
107 |
self.dreambooth_list = [os.path.basename(p) for p in dreambooth_list]
|
108 |
|
109 |
+
def update_dreambooth(self, dreambooth_dropdown, motion_module_dropdown=None):
|
110 |
self.selected_dreambooth = dreambooth_dropdown
|
111 |
|
112 |
dreambooth_dropdown = os.path.join(self.personalized_model_dir, dreambooth_dropdown)
|
|
|
117 |
converted_vae_checkpoint = convert_ldm_vae_checkpoint(dreambooth_state_dict, self.vae.config)
|
118 |
self.vae.load_state_dict(converted_vae_checkpoint)
|
119 |
|
120 |
+
del self.unet
|
121 |
+
self.unet = None
|
122 |
+
torch.cuda.empty_cache()
|
123 |
+
time.sleep(1)
|
|
|
|
|
|
|
|
|
124 |
converted_unet_checkpoint = convert_ldm_unet_checkpoint(dreambooth_state_dict, self.unet_model.config)
|
125 |
self.unet = copy.deepcopy(self.unet_model)
|
126 |
self.unet.load_state_dict(converted_unet_checkpoint, strict=False)
|
127 |
|
128 |
+
del self.text_encoder
|
129 |
+
self.text_encoder = None
|
130 |
+
torch.cuda.empty_cache()
|
131 |
+
time.sleep(1)
|
|
|
|
|
|
|
|
|
132 |
text_model = copy.deepcopy(self.text_model)
|
133 |
self.text_encoder = convert_ldm_clip_text_model(text_model, dreambooth_state_dict)
|
134 |
|
|
|
167 |
height_slider,
|
168 |
seed_textbox,
|
169 |
):
|
170 |
+
torch.cuda.empty_cache()
|
171 |
+
time.sleep(1)
|
172 |
+
|
173 |
if self.selected_motion_module != motion_module_dropdown: self.update_motion_module(motion_module_dropdown)
|
174 |
if self.selected_motion_module != motion_module_dropdown: self.update_motion_module_2(motion_module_dropdown)
|
175 |
if self.selected_dreambooth != dreambooth_dropdown: self.update_dreambooth(dreambooth_dropdown)
|
176 |
|
177 |
+
while self.text_encoder is None or self.unet is None:
|
178 |
+
self.update_dreambooth(dreambooth_dropdown, motion_module_dropdown)
|
179 |
+
|
180 |
if is_xformers_available(): self.unet.enable_xformers_memory_efficient_attention()
|
181 |
|
182 |
pipeline = MagicTimePipeline(
|
183 |
vae=self.vae, text_encoder=self.text_encoder, tokenizer=self.tokenizer, unet=self.unet,
|
184 |
scheduler=DDIMScheduler(**OmegaConf.to_container(self.inference_config.noise_scheduler_kwargs))
|
185 |
+
).to(device)
|
186 |
+
|
187 |
if int(seed_textbox) > 0: seed = int(seed_textbox)
|
188 |
else: seed = random.randint(1, 1e16)
|
189 |
torch.manual_seed(int(seed))
|
|
|
216 |
"seed": seed,
|
217 |
"dreambooth": dreambooth_dropdown,
|
218 |
}
|
219 |
+
|
220 |
+
torch.cuda.empty_cache()
|
221 |
+
time.sleep(1)
|
222 |
return gr.Video(value=save_sample_path), gr.Json(value=json_config)
|
223 |
|
224 |
+
controller = MagicTimeController()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
225 |
|
226 |
def ui():
|
227 |
with gr.Blocks(css=css) as demo:
|
|
|
242 |
dreambooth_dropdown = gr.Dropdown( label="DreamBooth Model", choices=controller.dreambooth_list, value=controller.dreambooth_list[0], interactive=True )
|
243 |
motion_module_dropdown = gr.Dropdown( label="Motion Module", choices=controller.motion_module_list, value=controller.motion_module_list[0], interactive=True )
|
244 |
|
|
|
|
|
|
|
245 |
prompt_textbox = gr.Textbox( label="Prompt", lines=3 )
|
246 |
negative_prompt_textbox = gr.Textbox( label="Negative Prompt", lines=3, value="worst quality, low quality, nsfw, logo")
|
247 |
|
|
|
274 |
|
275 |
return demo
|
276 |
|
|
|
277 |
if __name__ == "__main__":
|
278 |
demo = ui()
|
279 |
demo.queue(max_size=20)
|