Spaces:
Sleeping
Sleeping
liuhaotian
commited on
Commit
•
5c79044
1
Parent(s):
087de09
Fix OOM issue.
Browse files- app.py +29 -24
- gligen/task_grounded_generation.py +10 -6
app.py
CHANGED
@@ -2,7 +2,7 @@ import gradio as gr
|
|
2 |
import torch
|
3 |
import argparse
|
4 |
from omegaconf import OmegaConf
|
5 |
-
from gligen.task_grounded_generation import grounded_generation_box, load_ckpt
|
6 |
|
7 |
import json
|
8 |
import numpy as np
|
@@ -34,41 +34,46 @@ def parse_option():
|
|
34 |
args = parse_option()
|
35 |
|
36 |
|
37 |
-
def load_from_hf(repo_id, filename='diffusion_pytorch_model.bin'):
|
38 |
-
cache_file = hf_hub_download(repo_id=repo_id, filename=filename)
|
39 |
return torch.load(cache_file, map_location='cpu')
|
40 |
|
41 |
def load_ckpt_config_from_hf(modality):
|
42 |
-
ckpt = load_from_hf(
|
43 |
-
config = load_from_hf('gligen/
|
44 |
return ckpt, config
|
45 |
|
46 |
|
47 |
-
|
48 |
-
pretrained_ckpt_gligen, config = load_ckpt_config_from_hf(
|
49 |
config = OmegaConf.create( config["_content"] ) # config used in training
|
50 |
config.update( vars(args) )
|
51 |
-
config.model['params']['is_inpaint'] =
|
52 |
-
config.model['params']['is_style'] =
|
53 |
-
loaded_model_list = load_ckpt(config, pretrained_ckpt_gligen)
|
54 |
|
|
|
|
|
|
|
55 |
|
56 |
-
|
57 |
-
pretrained_ckpt_gligen_inpaint, config = load_ckpt_config_from_hf('gligen-inpainting-text-box')
|
58 |
-
config = OmegaConf.create( config["_content"] ) # config used in training
|
59 |
-
config.update( vars(args) )
|
60 |
-
config.model['params']['is_inpaint'] = True
|
61 |
-
config.model['params']['is_style'] = False
|
62 |
-
loaded_model_list_inpaint = load_ckpt(config, pretrained_ckpt_gligen_inpaint)
|
63 |
|
|
|
64 |
|
65 |
-
|
66 |
-
|
67 |
-
|
68 |
-
|
69 |
-
|
70 |
-
|
71 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
72 |
|
73 |
|
74 |
def load_clip_model():
|
|
|
2 |
import torch
|
3 |
import argparse
|
4 |
from omegaconf import OmegaConf
|
5 |
+
from gligen.task_grounded_generation import grounded_generation_box, load_ckpt, load_common_ckpt
|
6 |
|
7 |
import json
|
8 |
import numpy as np
|
|
|
34 |
args = parse_option()
|
35 |
|
36 |
|
37 |
+
def load_from_hf(repo_id, filename='diffusion_pytorch_model.bin', subfolder=None):
|
38 |
+
cache_file = hf_hub_download(repo_id=repo_id, filename=filename, subfolder=subfolder)
|
39 |
return torch.load(cache_file, map_location='cpu')
|
40 |
|
41 |
def load_ckpt_config_from_hf(modality):
|
42 |
+
ckpt = load_from_hf('gligen/demo_ckpts_legacy', filename=f'{modality}.pth', subfolder='model')
|
43 |
+
config = load_from_hf('gligen/demo_ckpts_legacy', filename=f'{modality}.pth', subfolder='config')
|
44 |
return ckpt, config
|
45 |
|
46 |
|
47 |
+
def ckpt_load_helper(modality, is_inpaint, is_style, common_instances=None):
|
48 |
+
pretrained_ckpt_gligen, config = load_ckpt_config_from_hf(modality)
|
49 |
config = OmegaConf.create( config["_content"] ) # config used in training
|
50 |
config.update( vars(args) )
|
51 |
+
config.model['params']['is_inpaint'] = is_inpaint
|
52 |
+
config.model['params']['is_style'] = is_style
|
|
|
53 |
|
54 |
+
if common_instances is None:
|
55 |
+
common_ckpt = load_from_hf('gligen/demo_ckpts_legacy', filename=f'common.pth', subfolder='model')
|
56 |
+
common_instances = load_common_ckpt(config, common_ckpt)
|
57 |
|
58 |
+
loaded_model_list = load_ckpt(config, pretrained_ckpt_gligen, common_instances)
|
|
|
|
|
|
|
|
|
|
|
|
|
59 |
|
60 |
+
return loaded_model_list, common_instances
|
61 |
|
62 |
+
|
63 |
+
loaded_model_list, common_instances = ckpt_load_helper(
|
64 |
+
'gligen-generation-text-box',
|
65 |
+
is_inpaint=False, is_style=False, common_instances=None
|
66 |
+
)
|
67 |
+
|
68 |
+
loaded_model_list_inpaint = ckpt_load_helper(
|
69 |
+
'gligen-inpainting-text-box',
|
70 |
+
is_inpaint=True, is_style=False, common_instances=common_instances
|
71 |
+
)[0]
|
72 |
+
|
73 |
+
loaded_model_list_style = ckpt_load_helper(
|
74 |
+
'gligen-generation-text-image-box',
|
75 |
+
is_inpaint=False, is_style=True, common_instances=common_instances
|
76 |
+
)[0]
|
77 |
|
78 |
|
79 |
def load_clip_model():
|
gligen/task_grounded_generation.py
CHANGED
@@ -65,21 +65,25 @@ def draw_box(img, locations):
|
|
65 |
draw.rectangle([box[0]*WW, box[1]*HH, box[2]*WW, box[3]*HH], outline =colors[bid % len(colors)], width=5)
|
66 |
return img
|
67 |
|
68 |
-
def
|
69 |
-
model = instantiate_from_config(config.model).to(device).eval()
|
70 |
autoencoder = instantiate_from_config(config.autoencoder).to(device).eval()
|
71 |
text_encoder = instantiate_from_config(config.text_encoder).to(device).eval()
|
72 |
diffusion = instantiate_from_config(config.diffusion).to(device)
|
73 |
|
74 |
-
autoencoder.load_state_dict(
|
75 |
-
text_encoder.load_state_dict(
|
76 |
-
diffusion.load_state_dict(
|
|
|
|
|
|
|
|
|
|
|
77 |
|
78 |
model.load_state_dict(state_dict['model'])
|
79 |
set_alpha_scale(model, config.alpha_scale)
|
80 |
print("ckpt is loaded")
|
81 |
|
82 |
-
return model
|
83 |
|
84 |
|
85 |
|
|
|
65 |
draw.rectangle([box[0]*WW, box[1]*HH, box[2]*WW, box[3]*HH], outline =colors[bid % len(colors)], width=5)
|
66 |
return img
|
67 |
|
68 |
+
def load_common_ckpt(config, common_ckpt):
|
|
|
69 |
autoencoder = instantiate_from_config(config.autoencoder).to(device).eval()
|
70 |
text_encoder = instantiate_from_config(config.text_encoder).to(device).eval()
|
71 |
diffusion = instantiate_from_config(config.diffusion).to(device)
|
72 |
|
73 |
+
autoencoder.load_state_dict( common_ckpt["autoencoder"] )
|
74 |
+
text_encoder.load_state_dict( common_ckpt["text_encoder"] )
|
75 |
+
diffusion.load_state_dict( common_ckpt["diffusion"] )
|
76 |
+
|
77 |
+
return [autoencoder, text_encoder, diffusion]
|
78 |
+
|
79 |
+
def load_ckpt(config, state_dict, common_instances):
|
80 |
+
model = instantiate_from_config(config.model).to(device).eval()
|
81 |
|
82 |
model.load_state_dict(state_dict['model'])
|
83 |
set_alpha_scale(model, config.alpha_scale)
|
84 |
print("ckpt is loaded")
|
85 |
|
86 |
+
return [model] + common_instances
|
87 |
|
88 |
|
89 |
|