Upload 5 files
Browse files- README (14).md +13 -0
- app (8).py +195 -0
- gitattributes (14) +35 -0
- kolors___init__.py +0 -0
- requirements (4).txt +8 -0
README (14).md
ADDED
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
title: Create_image
|
3 |
+
emoji: ๐ผ
|
4 |
+
colorFrom: purple
|
5 |
+
colorTo: red
|
6 |
+
sdk: gradio
|
7 |
+
sdk_version: 4.38.1
|
8 |
+
app_file: app.py
|
9 |
+
pinned: false
|
10 |
+
license: apache-2.0
|
11 |
+
---
|
12 |
+
|
13 |
+
Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
|
app (8).py
ADDED
@@ -0,0 +1,195 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import spaces
|
2 |
+
import random
|
3 |
+
import torch
|
4 |
+
from huggingface_hub import snapshot_download
|
5 |
+
from transformers import CLIPVisionModelWithProjection, CLIPImageProcessor
|
6 |
+
from kolors.pipelines import pipeline_stable_diffusion_xl_chatglm_256_ipadapter, pipeline_stable_diffusion_xl_chatglm_256
|
7 |
+
from kolors.models.modeling_chatglm import ChatGLMModel
|
8 |
+
from kolors.models.tokenization_chatglm import ChatGLMTokenizer
|
9 |
+
from kolors.models import unet_2d_condition
|
10 |
+
from diffusers import AutoencoderKL, EulerDiscreteScheduler, UNet2DConditionModel
|
11 |
+
import gradio as gr
|
12 |
+
import numpy as np
|
13 |
+
from huggingface_hub import InferenceClient
|
14 |
+
import os
|
15 |
+
|
16 |
+
device = "cuda"
|
17 |
+
ckpt_dir = snapshot_download(repo_id="Kwai-Kolors/Kolors")
|
18 |
+
ckpt_IPA_dir = snapshot_download(repo_id="Kwai-Kolors/Kolors-IP-Adapter-Plus")
|
19 |
+
|
20 |
+
text_encoder = ChatGLMModel.from_pretrained(f'{ckpt_dir}/text_encoder', torch_dtype=torch.float16).half().to(device)
|
21 |
+
tokenizer = ChatGLMTokenizer.from_pretrained(f'{ckpt_dir}/text_encoder')
|
22 |
+
vae = AutoencoderKL.from_pretrained(f"{ckpt_dir}/vae", revision=None).half().to(device)
|
23 |
+
scheduler = EulerDiscreteScheduler.from_pretrained(f"{ckpt_dir}/scheduler")
|
24 |
+
unet_t2i = UNet2DConditionModel.from_pretrained(f"{ckpt_dir}/unet", revision=None).half().to(device)
|
25 |
+
unet_i2i = unet_2d_condition.UNet2DConditionModel.from_pretrained(f"{ckpt_dir}/unet", revision=None).half().to(device)
|
26 |
+
image_encoder = CLIPVisionModelWithProjection.from_pretrained(f'{ckpt_IPA_dir}/image_encoder',ignore_mismatched_sizes=True).to(dtype=torch.float16, device=device)
|
27 |
+
ip_img_size = 336
|
28 |
+
clip_image_processor = CLIPImageProcessor(size=ip_img_size, crop_size=ip_img_size)
|
29 |
+
|
30 |
+
pipe_t2i = pipeline_stable_diffusion_xl_chatglm_256.StableDiffusionXLPipeline(
|
31 |
+
vae=vae,
|
32 |
+
text_encoder=text_encoder,
|
33 |
+
tokenizer=tokenizer,
|
34 |
+
unet=unet_t2i,
|
35 |
+
scheduler=scheduler,
|
36 |
+
force_zeros_for_empty_prompt=False
|
37 |
+
).to(device)
|
38 |
+
|
39 |
+
pipe_i2i = pipeline_stable_diffusion_xl_chatglm_256_ipadapter.StableDiffusionXLPipeline(
|
40 |
+
vae=vae,
|
41 |
+
text_encoder=text_encoder,
|
42 |
+
tokenizer=tokenizer,
|
43 |
+
unet=unet_i2i,
|
44 |
+
scheduler=scheduler,
|
45 |
+
image_encoder=image_encoder,
|
46 |
+
feature_extractor=clip_image_processor,
|
47 |
+
force_zeros_for_empty_prompt=False
|
48 |
+
).to(device)
|
49 |
+
|
50 |
+
if hasattr(pipe_i2i.unet, 'encoder_hid_proj'):
|
51 |
+
pipe_i2i.unet.text_encoder_hid_proj = pipe_i2i.unet.encoder_hid_proj
|
52 |
+
|
53 |
+
pipe_i2i.load_ip_adapter(f'{ckpt_IPA_dir}' , subfolder="", weight_name=["ip_adapter_plus_general.bin"])
|
54 |
+
|
55 |
+
MAX_SEED = np.iinfo(np.int32).max
|
56 |
+
MAX_IMAGE_SIZE = 1024
|
57 |
+
|
58 |
+
@spaces.GPU
|
59 |
+
def infer(prompt,
|
60 |
+
ip_adapter_image = None,
|
61 |
+
ip_adapter_scale = 0.5,
|
62 |
+
negative_prompt = "",
|
63 |
+
seed = 0,
|
64 |
+
randomize_seed = False,
|
65 |
+
width = 1024,
|
66 |
+
height = 1024,
|
67 |
+
guidance_scale = 5.0,
|
68 |
+
num_inference_steps = 25
|
69 |
+
):
|
70 |
+
if randomize_seed:
|
71 |
+
seed = random.randint(0, MAX_SEED)
|
72 |
+
generator = torch.Generator().manual_seed(seed)
|
73 |
+
|
74 |
+
if ip_adapter_image is None:
|
75 |
+
pipe_t2i.to(device)
|
76 |
+
image = pipe_t2i(
|
77 |
+
prompt = prompt,
|
78 |
+
negative_prompt = negative_prompt,
|
79 |
+
guidance_scale = guidance_scale,
|
80 |
+
num_inference_steps = num_inference_steps,
|
81 |
+
width = width,
|
82 |
+
height = height,
|
83 |
+
generator = generator
|
84 |
+
).images[0]
|
85 |
+
image.save("generated_image.jpg") # ํ์ผ ํ์ฅ์๋ฅผ .jpg๋ก ๋ณ๊ฒฝ
|
86 |
+
return image, "generated_image.jpg"
|
87 |
+
else:
|
88 |
+
pipe_i2i.to(device)
|
89 |
+
image_encoder.to(device)
|
90 |
+
pipe_i2i.image_encoder = image_encoder
|
91 |
+
pipe_i2i.set_ip_adapter_scale([ip_adapter_scale])
|
92 |
+
image = pipe_i2i(
|
93 |
+
prompt=prompt,
|
94 |
+
ip_adapter_image=[ip_adapter_image],
|
95 |
+
negative_prompt=negative_prompt,
|
96 |
+
height=height,
|
97 |
+
width=width,
|
98 |
+
num_inference_steps=num_inference_steps,
|
99 |
+
guidance_scale=guidance_scale,
|
100 |
+
num_images_per_prompt=1,
|
101 |
+
generator=generator
|
102 |
+
).images[0]
|
103 |
+
image.save("generated_image.jpg") # ํ์ผ ํ์ฅ์๋ฅผ .jpg๋ก ๋ณ๊ฒฝ
|
104 |
+
return image, "generated_image.jpg"
|
105 |
+
|
106 |
+
css="""
|
107 |
+
#col-left {
|
108 |
+
margin: 0 auto;
|
109 |
+
max-width: 600px;
|
110 |
+
}
|
111 |
+
#col-right {
|
112 |
+
margin: 0 auto;
|
113 |
+
max-width: 750px;
|
114 |
+
}
|
115 |
+
"""
|
116 |
+
|
117 |
+
with gr.Blocks(css=css) as Kolors:
|
118 |
+
with gr.Row():
|
119 |
+
with gr.Column(elem_id="col-left"):
|
120 |
+
with gr.Row():
|
121 |
+
generated_prompt = gr.Textbox(
|
122 |
+
label="ํ๋กฌํํธ ์
๋ ฅ",
|
123 |
+
placeholder="์ด๋ฏธ์ง ์์ฑ์ ์ฌ์ฉํ ํ๋กฌํํธ๋ฅผ ์
๋ ฅํ์ธ์",
|
124 |
+
lines=2
|
125 |
+
)
|
126 |
+
with gr.Row():
|
127 |
+
ip_adapter_image = gr.Image(label="Image Prompt (optional)", type="pil")
|
128 |
+
with gr.Row(visible=False): # Advanced Settings ์จ๊น
|
129 |
+
negative_prompt = gr.Textbox(
|
130 |
+
label="Negative prompt",
|
131 |
+
placeholder="Enter a negative prompt",
|
132 |
+
visible=True,
|
133 |
+
)
|
134 |
+
seed = gr.Slider(
|
135 |
+
label="Seed",
|
136 |
+
minimum=0,
|
137 |
+
maximum=MAX_SEED,
|
138 |
+
step=1,
|
139 |
+
value=0,
|
140 |
+
)
|
141 |
+
randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
|
142 |
+
with gr.Row():
|
143 |
+
width = gr.Slider(
|
144 |
+
label="Width",
|
145 |
+
minimum=256,
|
146 |
+
maximum=MAX_IMAGE_SIZE,
|
147 |
+
step=32,
|
148 |
+
value=1024,
|
149 |
+
)
|
150 |
+
height = gr.Slider(
|
151 |
+
label="Height",
|
152 |
+
minimum=256,
|
153 |
+
maximum=MAX_IMAGE_SIZE,
|
154 |
+
step=32,
|
155 |
+
value=1024,
|
156 |
+
)
|
157 |
+
with gr.Row():
|
158 |
+
guidance_scale = gr.Slider(
|
159 |
+
label="Guidance scale",
|
160 |
+
minimum=0.0,
|
161 |
+
maximum=10.0,
|
162 |
+
step=0.1,
|
163 |
+
value=5.0,
|
164 |
+
)
|
165 |
+
num_inference_steps = gr.Slider(
|
166 |
+
label="Number of inference steps",
|
167 |
+
minimum=10,
|
168 |
+
maximum=50,
|
169 |
+
step=1,
|
170 |
+
value=25,
|
171 |
+
)
|
172 |
+
with gr.Row():
|
173 |
+
ip_adapter_scale = gr.Slider(
|
174 |
+
label="Image influence scale",
|
175 |
+
info="Use 1 for creating variations",
|
176 |
+
minimum=0.0,
|
177 |
+
maximum=1.0,
|
178 |
+
step=0.05,
|
179 |
+
value=0.5,
|
180 |
+
)
|
181 |
+
with gr.Row():
|
182 |
+
run_button = gr.Button("Generate Image")
|
183 |
+
|
184 |
+
with gr.Column(elem_id="col-right"):
|
185 |
+
result = gr.Image(label="Result", show_label=False)
|
186 |
+
download_button = gr.File(label="Download Image")
|
187 |
+
|
188 |
+
# ์ด๋ฏธ์ง ์์ฑ ๋ฐ ๋ค์ด๋ก๋ ํ์ผ ๊ฒฝ๋ก ์ค์
|
189 |
+
run_button.click(
|
190 |
+
fn=infer,
|
191 |
+
inputs=[generated_prompt, ip_adapter_image, ip_adapter_scale, negative_prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps],
|
192 |
+
outputs=[result, download_button]
|
193 |
+
)
|
194 |
+
|
195 |
+
Kolors.queue().launch(debug=True)
|
gitattributes (14)
ADDED
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
*.7z filter=lfs diff=lfs merge=lfs -text
|
2 |
+
*.arrow filter=lfs diff=lfs merge=lfs -text
|
3 |
+
*.bin filter=lfs diff=lfs merge=lfs -text
|
4 |
+
*.bz2 filter=lfs diff=lfs merge=lfs -text
|
5 |
+
*.ckpt filter=lfs diff=lfs merge=lfs -text
|
6 |
+
*.ftz filter=lfs diff=lfs merge=lfs -text
|
7 |
+
*.gz filter=lfs diff=lfs merge=lfs -text
|
8 |
+
*.h5 filter=lfs diff=lfs merge=lfs -text
|
9 |
+
*.joblib filter=lfs diff=lfs merge=lfs -text
|
10 |
+
*.lfs.* filter=lfs diff=lfs merge=lfs -text
|
11 |
+
*.mlmodel filter=lfs diff=lfs merge=lfs -text
|
12 |
+
*.model filter=lfs diff=lfs merge=lfs -text
|
13 |
+
*.msgpack filter=lfs diff=lfs merge=lfs -text
|
14 |
+
*.npy filter=lfs diff=lfs merge=lfs -text
|
15 |
+
*.npz filter=lfs diff=lfs merge=lfs -text
|
16 |
+
*.onnx filter=lfs diff=lfs merge=lfs -text
|
17 |
+
*.ot filter=lfs diff=lfs merge=lfs -text
|
18 |
+
*.parquet filter=lfs diff=lfs merge=lfs -text
|
19 |
+
*.pb filter=lfs diff=lfs merge=lfs -text
|
20 |
+
*.pickle filter=lfs diff=lfs merge=lfs -text
|
21 |
+
*.pkl filter=lfs diff=lfs merge=lfs -text
|
22 |
+
*.pt filter=lfs diff=lfs merge=lfs -text
|
23 |
+
*.pth filter=lfs diff=lfs merge=lfs -text
|
24 |
+
*.rar filter=lfs diff=lfs merge=lfs -text
|
25 |
+
*.safetensors filter=lfs diff=lfs merge=lfs -text
|
26 |
+
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
27 |
+
*.tar.* filter=lfs diff=lfs merge=lfs -text
|
28 |
+
*.tar filter=lfs diff=lfs merge=lfs -text
|
29 |
+
*.tflite filter=lfs diff=lfs merge=lfs -text
|
30 |
+
*.tgz filter=lfs diff=lfs merge=lfs -text
|
31 |
+
*.wasm filter=lfs diff=lfs merge=lfs -text
|
32 |
+
*.xz filter=lfs diff=lfs merge=lfs -text
|
33 |
+
*.zip filter=lfs diff=lfs merge=lfs -text
|
34 |
+
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
+
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
kolors___init__.py
ADDED
File without changes
|
requirements (4).txt
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
accelerate==0.27.2
|
2 |
+
diffusers==0.28.2
|
3 |
+
invisible_watermark==0.2.0
|
4 |
+
torch==2.2.0
|
5 |
+
transformers==4.42.4
|
6 |
+
sentencepiece==0.1.99
|
7 |
+
huggingface_hub
|
8 |
+
gradio
|