fffiloni hysts HF staff commited on
Commit
a1a09dc
0 Parent(s):

Duplicate from hysts/ControlNet

Browse files

Co-authored-by: hysts <hysts@users.noreply.huggingface.co>

.gitattributes ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ *.tflite filter=lfs diff=lfs merge=lfs -text
29
+ *.tgz filter=lfs diff=lfs merge=lfs -text
30
+ *.wasm filter=lfs diff=lfs merge=lfs -text
31
+ *.xz filter=lfs diff=lfs merge=lfs -text
32
+ *.zip filter=lfs diff=lfs merge=lfs -text
33
+ *.zst filter=lfs diff=lfs merge=lfs -text
34
+ *tfevents* filter=lfs diff=lfs merge=lfs -text
.gitignore ADDED
@@ -0,0 +1,162 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ models/
2
+
3
+ # Byte-compiled / optimized / DLL files
4
+ __pycache__/
5
+ *.py[cod]
6
+ *$py.class
7
+
8
+ # C extensions
9
+ *.so
10
+
11
+ # Distribution / packaging
12
+ .Python
13
+ build/
14
+ develop-eggs/
15
+ dist/
16
+ downloads/
17
+ eggs/
18
+ .eggs/
19
+ lib/
20
+ lib64/
21
+ parts/
22
+ sdist/
23
+ var/
24
+ wheels/
25
+ share/python-wheels/
26
+ *.egg-info/
27
+ .installed.cfg
28
+ *.egg
29
+ MANIFEST
30
+
31
+ # PyInstaller
32
+ # Usually these files are written by a python script from a template
33
+ # before PyInstaller builds the exe, so as to inject date/other infos into it.
34
+ *.manifest
35
+ *.spec
36
+
37
+ # Installer logs
38
+ pip-log.txt
39
+ pip-delete-this-directory.txt
40
+
41
+ # Unit test / coverage reports
42
+ htmlcov/
43
+ .tox/
44
+ .nox/
45
+ .coverage
46
+ .coverage.*
47
+ .cache
48
+ nosetests.xml
49
+ coverage.xml
50
+ *.cover
51
+ *.py,cover
52
+ .hypothesis/
53
+ .pytest_cache/
54
+ cover/
55
+
56
+ # Translations
57
+ *.mo
58
+ *.pot
59
+
60
+ # Django stuff:
61
+ *.log
62
+ local_settings.py
63
+ db.sqlite3
64
+ db.sqlite3-journal
65
+
66
+ # Flask stuff:
67
+ instance/
68
+ .webassets-cache
69
+
70
+ # Scrapy stuff:
71
+ .scrapy
72
+
73
+ # Sphinx documentation
74
+ docs/_build/
75
+
76
+ # PyBuilder
77
+ .pybuilder/
78
+ target/
79
+
80
+ # Jupyter Notebook
81
+ .ipynb_checkpoints
82
+
83
+ # IPython
84
+ profile_default/
85
+ ipython_config.py
86
+
87
+ # pyenv
88
+ # For a library or package, you might want to ignore these files since the code is
89
+ # intended to run in multiple environments; otherwise, check them in:
90
+ # .python-version
91
+
92
+ # pipenv
93
+ # According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
94
+ # However, in case of collaboration, if having platform-specific dependencies or dependencies
95
+ # having no cross-platform support, pipenv may install dependencies that don't work, or not
96
+ # install all needed dependencies.
97
+ #Pipfile.lock
98
+
99
+ # poetry
100
+ # Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
101
+ # This is especially recommended for binary packages to ensure reproducibility, and is more
102
+ # commonly ignored for libraries.
103
+ # https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
104
+ #poetry.lock
105
+
106
+ # pdm
107
+ # Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
108
+ #pdm.lock
109
+ # pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
110
+ # in version control.
111
+ # https://pdm.fming.dev/#use-with-ide
112
+ .pdm.toml
113
+
114
+ # PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
115
+ __pypackages__/
116
+
117
+ # Celery stuff
118
+ celerybeat-schedule
119
+ celerybeat.pid
120
+
121
+ # SageMath parsed files
122
+ *.sage.py
123
+
124
+ # Environments
125
+ .env
126
+ .venv
127
+ env/
128
+ venv/
129
+ ENV/
130
+ env.bak/
131
+ venv.bak/
132
+
133
+ # Spyder project settings
134
+ .spyderproject
135
+ .spyproject
136
+
137
+ # Rope project settings
138
+ .ropeproject
139
+
140
+ # mkdocs documentation
141
+ /site
142
+
143
+ # mypy
144
+ .mypy_cache/
145
+ .dmypy.json
146
+ dmypy.json
147
+
148
+ # Pyre type checker
149
+ .pyre/
150
+
151
+ # pytype static type analyzer
152
+ .pytype/
153
+
154
+ # Cython debug symbols
155
+ cython_debug/
156
+
157
+ # PyCharm
158
+ # JetBrains specific template is maintained in a separate JetBrains.gitignore that can
159
+ # be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
160
+ # and can be added to the global gitignore or merged into this file. For a more nuclear
161
+ # option (not recommended) you can uncomment the following to ignore the entire idea folder.
162
+ #.idea/
.gitmodules ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ [submodule "ControlNet"]
2
+ path = ControlNet
3
+ url = https://github.com/lllyasviel/ControlNet
.pre-commit-config.yaml ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ exclude: patch
2
+ repos:
3
+ - repo: https://github.com/pre-commit/pre-commit-hooks
4
+ rev: v4.2.0
5
+ hooks:
6
+ - id: check-executables-have-shebangs
7
+ - id: check-json
8
+ - id: check-merge-conflict
9
+ - id: check-shebang-scripts-are-executable
10
+ - id: check-toml
11
+ - id: check-yaml
12
+ - id: double-quote-string-fixer
13
+ - id: end-of-file-fixer
14
+ - id: mixed-line-ending
15
+ args: ['--fix=lf']
16
+ - id: requirements-txt-fixer
17
+ - id: trailing-whitespace
18
+ - repo: https://github.com/myint/docformatter
19
+ rev: v1.4
20
+ hooks:
21
+ - id: docformatter
22
+ args: ['--in-place']
23
+ - repo: https://github.com/pycqa/isort
24
+ rev: 5.12.0
25
+ hooks:
26
+ - id: isort
27
+ - repo: https://github.com/pre-commit/mirrors-mypy
28
+ rev: v0.991
29
+ hooks:
30
+ - id: mypy
31
+ args: ['--ignore-missing-imports']
32
+ additional_dependencies: ['types-python-slugify']
33
+ - repo: https://github.com/google/yapf
34
+ rev: v0.32.0
35
+ hooks:
36
+ - id: yapf
37
+ args: ['--parallel', '--in-place']
.style.yapf ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ [style]
2
+ based_on_style = pep8
3
+ blank_line_before_nested_class_or_def = false
4
+ spaces_before_comment = 2
5
+ split_before_logical_operator = true
ControlNet ADDED
@@ -0,0 +1 @@
 
 
1
+ Subproject commit f4748e3630d8141d7765e2bd9b1e348f47847707
LICENSE.ControlNet ADDED
@@ -0,0 +1,201 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Apache License
2
+ Version 2.0, January 2004
3
+ http://www.apache.org/licenses/
4
+
5
+ TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
6
+
7
+ 1. Definitions.
8
+
9
+ "License" shall mean the terms and conditions for use, reproduction,
10
+ and distribution as defined by Sections 1 through 9 of this document.
11
+
12
+ "Licensor" shall mean the copyright owner or entity authorized by
13
+ the copyright owner that is granting the License.
14
+
15
+ "Legal Entity" shall mean the union of the acting entity and all
16
+ other entities that control, are controlled by, or are under common
17
+ control with that entity. For the purposes of this definition,
18
+ "control" means (i) the power, direct or indirect, to cause the
19
+ direction or management of such entity, whether by contract or
20
+ otherwise, or (ii) ownership of fifty percent (50%) or more of the
21
+ outstanding shares, or (iii) beneficial ownership of such entity.
22
+
23
+ "You" (or "Your") shall mean an individual or Legal Entity
24
+ exercising permissions granted by this License.
25
+
26
+ "Source" form shall mean the preferred form for making modifications,
27
+ including but not limited to software source code, documentation
28
+ source, and configuration files.
29
+
30
+ "Object" form shall mean any form resulting from mechanical
31
+ transformation or translation of a Source form, including but
32
+ not limited to compiled object code, generated documentation,
33
+ and conversions to other media types.
34
+
35
+ "Work" shall mean the work of authorship, whether in Source or
36
+ Object form, made available under the License, as indicated by a
37
+ copyright notice that is included in or attached to the work
38
+ (an example is provided in the Appendix below).
39
+
40
+ "Derivative Works" shall mean any work, whether in Source or Object
41
+ form, that is based on (or derived from) the Work and for which the
42
+ editorial revisions, annotations, elaborations, or other modifications
43
+ represent, as a whole, an original work of authorship. For the purposes
44
+ of this License, Derivative Works shall not include works that remain
45
+ separable from, or merely link (or bind by name) to the interfaces of,
46
+ the Work and Derivative Works thereof.
47
+
48
+ "Contribution" shall mean any work of authorship, including
49
+ the original version of the Work and any modifications or additions
50
+ to that Work or Derivative Works thereof, that is intentionally
51
+ submitted to Licensor for inclusion in the Work by the copyright owner
52
+ or by an individual or Legal Entity authorized to submit on behalf of
53
+ the copyright owner. For the purposes of this definition, "submitted"
54
+ means any form of electronic, verbal, or written communication sent
55
+ to the Licensor or its representatives, including but not limited to
56
+ communication on electronic mailing lists, source code control systems,
57
+ and issue tracking systems that are managed by, or on behalf of, the
58
+ Licensor for the purpose of discussing and improving the Work, but
59
+ excluding communication that is conspicuously marked or otherwise
60
+ designated in writing by the copyright owner as "Not a Contribution."
61
+
62
+ "Contributor" shall mean Licensor and any individual or Legal Entity
63
+ on behalf of whom a Contribution has been received by Licensor and
64
+ subsequently incorporated within the Work.
65
+
66
+ 2. Grant of Copyright License. Subject to the terms and conditions of
67
+ this License, each Contributor hereby grants to You a perpetual,
68
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
69
+ copyright license to reproduce, prepare Derivative Works of,
70
+ publicly display, publicly perform, sublicense, and distribute the
71
+ Work and such Derivative Works in Source or Object form.
72
+
73
+ 3. Grant of Patent License. Subject to the terms and conditions of
74
+ this License, each Contributor hereby grants to You a perpetual,
75
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
76
+ (except as stated in this section) patent license to make, have made,
77
+ use, offer to sell, sell, import, and otherwise transfer the Work,
78
+ where such license applies only to those patent claims licensable
79
+ by such Contributor that are necessarily infringed by their
80
+ Contribution(s) alone or by combination of their Contribution(s)
81
+ with the Work to which such Contribution(s) was submitted. If You
82
+ institute patent litigation against any entity (including a
83
+ cross-claim or counterclaim in a lawsuit) alleging that the Work
84
+ or a Contribution incorporated within the Work constitutes direct
85
+ or contributory patent infringement, then any patent licenses
86
+ granted to You under this License for that Work shall terminate
87
+ as of the date such litigation is filed.
88
+
89
+ 4. Redistribution. You may reproduce and distribute copies of the
90
+ Work or Derivative Works thereof in any medium, with or without
91
+ modifications, and in Source or Object form, provided that You
92
+ meet the following conditions:
93
+
94
+ (a) You must give any other recipients of the Work or
95
+ Derivative Works a copy of this License; and
96
+
97
+ (b) You must cause any modified files to carry prominent notices
98
+ stating that You changed the files; and
99
+
100
+ (c) You must retain, in the Source form of any Derivative Works
101
+ that You distribute, all copyright, patent, trademark, and
102
+ attribution notices from the Source form of the Work,
103
+ excluding those notices that do not pertain to any part of
104
+ the Derivative Works; and
105
+
106
+ (d) If the Work includes a "NOTICE" text file as part of its
107
+ distribution, then any Derivative Works that You distribute must
108
+ include a readable copy of the attribution notices contained
109
+ within such NOTICE file, excluding those notices that do not
110
+ pertain to any part of the Derivative Works, in at least one
111
+ of the following places: within a NOTICE text file distributed
112
+ as part of the Derivative Works; within the Source form or
113
+ documentation, if provided along with the Derivative Works; or,
114
+ within a display generated by the Derivative Works, if and
115
+ wherever such third-party notices normally appear. The contents
116
+ of the NOTICE file are for informational purposes only and
117
+ do not modify the License. You may add Your own attribution
118
+ notices within Derivative Works that You distribute, alongside
119
+ or as an addendum to the NOTICE text from the Work, provided
120
+ that such additional attribution notices cannot be construed
121
+ as modifying the License.
122
+
123
+ You may add Your own copyright statement to Your modifications and
124
+ may provide additional or different license terms and conditions
125
+ for use, reproduction, or distribution of Your modifications, or
126
+ for any such Derivative Works as a whole, provided Your use,
127
+ reproduction, and distribution of the Work otherwise complies with
128
+ the conditions stated in this License.
129
+
130
+ 5. Submission of Contributions. Unless You explicitly state otherwise,
131
+ any Contribution intentionally submitted for inclusion in the Work
132
+ by You to the Licensor shall be under the terms and conditions of
133
+ this License, without any additional terms or conditions.
134
+ Notwithstanding the above, nothing herein shall supersede or modify
135
+ the terms of any separate license agreement you may have executed
136
+ with Licensor regarding such Contributions.
137
+
138
+ 6. Trademarks. This License does not grant permission to use the trade
139
+ names, trademarks, service marks, or product names of the Licensor,
140
+ except as required for reasonable and customary use in describing the
141
+ origin of the Work and reproducing the content of the NOTICE file.
142
+
143
+ 7. Disclaimer of Warranty. Unless required by applicable law or
144
+ agreed to in writing, Licensor provides the Work (and each
145
+ Contributor provides its Contributions) on an "AS IS" BASIS,
146
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
147
+ implied, including, without limitation, any warranties or conditions
148
+ of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
149
+ PARTICULAR PURPOSE. You are solely responsible for determining the
150
+ appropriateness of using or redistributing the Work and assume any
151
+ risks associated with Your exercise of permissions under this License.
152
+
153
+ 8. Limitation of Liability. In no event and under no legal theory,
154
+ whether in tort (including negligence), contract, or otherwise,
155
+ unless required by applicable law (such as deliberate and grossly
156
+ negligent acts) or agreed to in writing, shall any Contributor be
157
+ liable to You for damages, including any direct, indirect, special,
158
+ incidental, or consequential damages of any character arising as a
159
+ result of this License or out of the use or inability to use the
160
+ Work (including but not limited to damages for loss of goodwill,
161
+ work stoppage, computer failure or malfunction, or any and all
162
+ other commercial damages or losses), even if such Contributor
163
+ has been advised of the possibility of such damages.
164
+
165
+ 9. Accepting Warranty or Additional Liability. While redistributing
166
+ the Work or Derivative Works thereof, You may choose to offer,
167
+ and charge a fee for, acceptance of support, warranty, indemnity,
168
+ or other liability obligations and/or rights consistent with this
169
+ License. However, in accepting such obligations, You may act only
170
+ on Your own behalf and on Your sole responsibility, not on behalf
171
+ of any other Contributor, and only if You agree to indemnify,
172
+ defend, and hold each Contributor harmless for any liability
173
+ incurred by, or claims asserted against, such Contributor by reason
174
+ of your accepting any such warranty or additional liability.
175
+
176
+ END OF TERMS AND CONDITIONS
177
+
178
+ APPENDIX: How to apply the Apache License to your work.
179
+
180
+ To apply the Apache License to your work, attach the following
181
+ boilerplate notice, with the fields enclosed by brackets "[]"
182
+ replaced with your own identifying information. (Don't include
183
+ the brackets!) The text should be enclosed in the appropriate
184
+ comment syntax for the file format. We also recommend that a
185
+ file or class name and description of purpose be included on the
186
+ same "printed page" as the copyright notice for easier
187
+ identification within third-party archives.
188
+
189
+ Copyright [yyyy] [name of copyright owner]
190
+
191
+ Licensed under the Apache License, Version 2.0 (the "License");
192
+ you may not use this file except in compliance with the License.
193
+ You may obtain a copy of the License at
194
+
195
+ http://www.apache.org/licenses/LICENSE-2.0
196
+
197
+ Unless required by applicable law or agreed to in writing, software
198
+ distributed under the License is distributed on an "AS IS" BASIS,
199
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
200
+ See the License for the specific language governing permissions and
201
+ limitations under the License.
README.md ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ title: ControlNet
3
+ emoji: 🌖
4
+ colorFrom: pink
5
+ colorTo: blue
6
+ sdk: gradio
7
+ sdk_version: 3.18.0
8
+ python_version: 3.10.9
9
+ app_file: app.py
10
+ pinned: false
11
+ duplicated_from: hysts/ControlNet
12
+ ---
13
+
14
+ Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
app.py ADDED
@@ -0,0 +1,88 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python
2
+
3
+ from __future__ import annotations
4
+
5
+ import os
6
+ import pathlib
7
+ import shlex
8
+ import subprocess
9
+
10
+ import gradio as gr
11
+
12
+ if os.getenv('SYSTEM') == 'spaces':
13
+ with open('patch') as f:
14
+ subprocess.run(shlex.split('patch -p1'), stdin=f, cwd='ControlNet')
15
+
16
+ base_url = 'https://huggingface.co/lllyasviel/ControlNet/resolve/main/annotator/ckpts/'
17
+ names = [
18
+ 'body_pose_model.pth',
19
+ 'dpt_hybrid-midas-501f0c75.pt',
20
+ 'hand_pose_model.pth',
21
+ 'mlsd_large_512_fp32.pth',
22
+ 'mlsd_tiny_512_fp32.pth',
23
+ 'network-bsds500.pth',
24
+ 'upernet_global_small.pth',
25
+ ]
26
+ for name in names:
27
+ command = f'wget https://huggingface.co/lllyasviel/ControlNet/resolve/main/annotator/ckpts/{name} -O {name}'
28
+ out_path = pathlib.Path(f'ControlNet/annotator/ckpts/{name}')
29
+ if out_path.exists():
30
+ continue
31
+ subprocess.run(shlex.split(command), cwd='ControlNet/annotator/ckpts/')
32
+
33
+ from gradio_canny2image import create_demo as create_demo_canny
34
+ from gradio_depth2image import create_demo as create_demo_depth
35
+ from gradio_fake_scribble2image import create_demo as create_demo_fake_scribble
36
+ from gradio_hed2image import create_demo as create_demo_hed
37
+ from gradio_hough2image import create_demo as create_demo_hough
38
+ from gradio_normal2image import create_demo as create_demo_normal
39
+ from gradio_pose2image import create_demo as create_demo_pose
40
+ from gradio_scribble2image import create_demo as create_demo_scribble
41
+ from gradio_scribble2image_interactive import \
42
+ create_demo as create_demo_scribble_interactive
43
+ from gradio_seg2image import create_demo as create_demo_seg
44
+ from model import Model
45
+
46
+ MAX_IMAGES = 1
47
+ DESCRIPTION = '''# ControlNet
48
+
49
+ This is an unofficial demo for [https://github.com/lllyasviel/ControlNet](https://github.com/lllyasviel/ControlNet).
50
+
51
+ If you are interested in trying out other base models, check out [this Space](https://huggingface.co/spaces/hysts/ControlNet-with-other-models) as well.
52
+ '''
53
+ if (SPACE_ID := os.getenv('SPACE_ID')) is not None:
54
+ DESCRIPTION += f'''<p>For faster inference without waiting in queue, you may duplicate the space and upgrade to GPU in settings.<br/>
55
+ <a href="https://huggingface.co/spaces/{SPACE_ID}?duplicate=true">
56
+ <img style="margin-top: 0em; margin-bottom: 0em" src="https://bit.ly/3gLdBN6" alt="Duplicate Space"></a>
57
+ <p/>
58
+ '''
59
+
60
+ model = Model()
61
+
62
+ with gr.Blocks(css='style.css') as demo:
63
+ gr.Markdown(DESCRIPTION)
64
+ with gr.Tabs():
65
+ with gr.TabItem('Canny'):
66
+ create_demo_canny(model.process_canny, max_images=MAX_IMAGES)
67
+ with gr.TabItem('Hough'):
68
+ create_demo_hough(model.process_hough, max_images=MAX_IMAGES)
69
+ with gr.TabItem('HED'):
70
+ create_demo_hed(model.process_hed, max_images=MAX_IMAGES)
71
+ with gr.TabItem('Scribble'):
72
+ create_demo_scribble(model.process_scribble, max_images=MAX_IMAGES)
73
+ with gr.TabItem('Scribble Interactive'):
74
+ create_demo_scribble_interactive(
75
+ model.process_scribble_interactive, max_images=MAX_IMAGES)
76
+ with gr.TabItem('Fake Scribble'):
77
+ create_demo_fake_scribble(model.process_fake_scribble,
78
+ max_images=MAX_IMAGES)
79
+ with gr.TabItem('Pose'):
80
+ create_demo_pose(model.process_pose, max_images=MAX_IMAGES)
81
+ with gr.TabItem('Segmentation'):
82
+ create_demo_seg(model.process_seg, max_images=MAX_IMAGES)
83
+ with gr.TabItem('Depth'):
84
+ create_demo_depth(model.process_depth, max_images=MAX_IMAGES)
85
+ with gr.TabItem('Normal map'):
86
+ create_demo_normal(model.process_normal, max_images=MAX_IMAGES)
87
+
88
+ demo.queue(api_open=False).launch()
gradio_canny2image.py ADDED
@@ -0,0 +1,75 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # This file is adapted from https://github.com/lllyasviel/ControlNet/blob/f4748e3630d8141d7765e2bd9b1e348f47847707/gradio_canny2image.py
2
+ # The original license file is LICENSE.ControlNet in this repo.
3
+ import gradio as gr
4
+
5
+
6
+ def create_demo(process, max_images=12):
7
+ with gr.Blocks() as demo:
8
+ with gr.Row():
9
+ gr.Markdown('## Control Stable Diffusion with Canny Edge Maps')
10
+ with gr.Row():
11
+ with gr.Column():
12
+ input_image = gr.Image(source='upload', type='numpy')
13
+ prompt = gr.Textbox(label='Prompt')
14
+ run_button = gr.Button(label='Run')
15
+ with gr.Accordion('Advanced options', open=False):
16
+ num_samples = gr.Slider(label='Images',
17
+ minimum=1,
18
+ maximum=max_images,
19
+ value=1,
20
+ step=1)
21
+ image_resolution = gr.Slider(label='Image Resolution',
22
+ minimum=256,
23
+ maximum=768,
24
+ value=512,
25
+ step=256)
26
+ low_threshold = gr.Slider(label='Canny low threshold',
27
+ minimum=1,
28
+ maximum=255,
29
+ value=100,
30
+ step=1)
31
+ high_threshold = gr.Slider(label='Canny high threshold',
32
+ minimum=1,
33
+ maximum=255,
34
+ value=200,
35
+ step=1)
36
+ ddim_steps = gr.Slider(label='Steps',
37
+ minimum=1,
38
+ maximum=100,
39
+ value=20,
40
+ step=1)
41
+ scale = gr.Slider(label='Guidance Scale',
42
+ minimum=0.1,
43
+ maximum=30.0,
44
+ value=9.0,
45
+ step=0.1)
46
+ seed = gr.Slider(label='Seed',
47
+ minimum=-1,
48
+ maximum=2147483647,
49
+ step=1,
50
+ randomize=True,
51
+ queue=False)
52
+ eta = gr.Number(label='eta (DDIM)', value=0.0)
53
+ a_prompt = gr.Textbox(
54
+ label='Added Prompt',
55
+ value='best quality, extremely detailed')
56
+ n_prompt = gr.Textbox(
57
+ label='Negative Prompt',
58
+ value=
59
+ 'longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality'
60
+ )
61
+ with gr.Column():
62
+ result_gallery = gr.Gallery(label='Output',
63
+ show_label=False,
64
+ elem_id='gallery').style(
65
+ grid=2, height='auto')
66
+ ips = [
67
+ input_image, prompt, a_prompt, n_prompt, num_samples,
68
+ image_resolution, ddim_steps, scale, seed, eta, low_threshold,
69
+ high_threshold
70
+ ]
71
+ run_button.click(fn=process,
72
+ inputs=ips,
73
+ outputs=[result_gallery],
74
+ api_name='canny')
75
+ return demo
gradio_depth2image.py ADDED
@@ -0,0 +1,69 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # This file is adapted from https://github.com/lllyasviel/ControlNet/blob/f4748e3630d8141d7765e2bd9b1e348f47847707/gradio_depth2image.py
2
+ # The original license file is LICENSE.ControlNet in this repo.
3
+ import gradio as gr
4
+
5
+
6
+ def create_demo(process, max_images=12):
7
+ with gr.Blocks() as demo:
8
+ with gr.Row():
9
+ gr.Markdown('## Control Stable Diffusion with Depth Maps')
10
+ with gr.Row():
11
+ with gr.Column():
12
+ input_image = gr.Image(source='upload', type='numpy')
13
+ prompt = gr.Textbox(label='Prompt')
14
+ run_button = gr.Button(label='Run')
15
+ with gr.Accordion('Advanced options', open=False):
16
+ num_samples = gr.Slider(label='Images',
17
+ minimum=1,
18
+ maximum=max_images,
19
+ value=1,
20
+ step=1)
21
+ image_resolution = gr.Slider(label='Image Resolution',
22
+ minimum=256,
23
+ maximum=768,
24
+ value=512,
25
+ step=256)
26
+ detect_resolution = gr.Slider(label='Depth Resolution',
27
+ minimum=128,
28
+ maximum=1024,
29
+ value=384,
30
+ step=1)
31
+ ddim_steps = gr.Slider(label='Steps',
32
+ minimum=1,
33
+ maximum=100,
34
+ value=20,
35
+ step=1)
36
+ scale = gr.Slider(label='Guidance Scale',
37
+ minimum=0.1,
38
+ maximum=30.0,
39
+ value=9.0,
40
+ step=0.1)
41
+ seed = gr.Slider(label='Seed',
42
+ minimum=-1,
43
+ maximum=2147483647,
44
+ step=1,
45
+ randomize=True,
46
+ queue=False)
47
+ eta = gr.Number(label='eta (DDIM)', value=0.0)
48
+ a_prompt = gr.Textbox(
49
+ label='Added Prompt',
50
+ value='best quality, extremely detailed')
51
+ n_prompt = gr.Textbox(
52
+ label='Negative Prompt',
53
+ value=
54
+ 'longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality'
55
+ )
56
+ with gr.Column():
57
+ result_gallery = gr.Gallery(label='Output',
58
+ show_label=False,
59
+ elem_id='gallery').style(
60
+ grid=2, height='auto')
61
+ ips = [
62
+ input_image, prompt, a_prompt, n_prompt, num_samples,
63
+ image_resolution, detect_resolution, ddim_steps, scale, seed, eta
64
+ ]
65
+ run_button.click(fn=process,
66
+ inputs=ips,
67
+ outputs=[result_gallery],
68
+ api_name='depth')
69
+ return demo
gradio_fake_scribble2image.py ADDED
@@ -0,0 +1,69 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # This file is adapted from https://github.com/lllyasviel/ControlNet/blob/f4748e3630d8141d7765e2bd9b1e348f47847707/gradio_fake_scribble2image.py
2
+ # The original license file is LICENSE.ControlNet in this repo.
3
+ import gradio as gr
4
+
5
+
6
+ def create_demo(process, max_images=12):
7
+ with gr.Blocks() as demo:
8
+ with gr.Row():
9
+ gr.Markdown('## Control Stable Diffusion with Fake Scribble Maps')
10
+ with gr.Row():
11
+ with gr.Column():
12
+ input_image = gr.Image(source='upload', type='numpy')
13
+ prompt = gr.Textbox(label='Prompt')
14
+ run_button = gr.Button(label='Run')
15
+ with gr.Accordion('Advanced options', open=False):
16
+ num_samples = gr.Slider(label='Images',
17
+ minimum=1,
18
+ maximum=max_images,
19
+ value=1,
20
+ step=1)
21
+ image_resolution = gr.Slider(label='Image Resolution',
22
+ minimum=256,
23
+ maximum=768,
24
+ value=512,
25
+ step=256)
26
+ detect_resolution = gr.Slider(label='HED Resolution',
27
+ minimum=128,
28
+ maximum=1024,
29
+ value=512,
30
+ step=1)
31
+ ddim_steps = gr.Slider(label='Steps',
32
+ minimum=1,
33
+ maximum=100,
34
+ value=20,
35
+ step=1)
36
+ scale = gr.Slider(label='Guidance Scale',
37
+ minimum=0.1,
38
+ maximum=30.0,
39
+ value=9.0,
40
+ step=0.1)
41
+ seed = gr.Slider(label='Seed',
42
+ minimum=-1,
43
+ maximum=2147483647,
44
+ step=1,
45
+ randomize=True,
46
+ queue=False)
47
+ eta = gr.Number(label='eta (DDIM)', value=0.0)
48
+ a_prompt = gr.Textbox(
49
+ label='Added Prompt',
50
+ value='best quality, extremely detailed')
51
+ n_prompt = gr.Textbox(
52
+ label='Negative Prompt',
53
+ value=
54
+ 'longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality'
55
+ )
56
+ with gr.Column():
57
+ result_gallery = gr.Gallery(label='Output',
58
+ show_label=False,
59
+ elem_id='gallery').style(
60
+ grid=2, height='auto')
61
+ ips = [
62
+ input_image, prompt, a_prompt, n_prompt, num_samples,
63
+ image_resolution, detect_resolution, ddim_steps, scale, seed, eta
64
+ ]
65
+ run_button.click(fn=process,
66
+ inputs=ips,
67
+ outputs=[result_gallery],
68
+ api_name='fake_scribble')
69
+ return demo
gradio_hed2image.py ADDED
@@ -0,0 +1,69 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # This file is adapted from https://github.com/lllyasviel/ControlNet/blob/f4748e3630d8141d7765e2bd9b1e348f47847707/gradio_hed2image.py
2
+ # The original license file is LICENSE.ControlNet in this repo.
3
+ import gradio as gr
4
+
5
+
6
+ def create_demo(process, max_images=12):
7
+ with gr.Blocks() as demo:
8
+ with gr.Row():
9
+ gr.Markdown('## Control Stable Diffusion with HED Maps')
10
+ with gr.Row():
11
+ with gr.Column():
12
+ input_image = gr.Image(source='upload', type='numpy')
13
+ prompt = gr.Textbox(label='Prompt')
14
+ run_button = gr.Button(label='Run')
15
+ with gr.Accordion('Advanced options', open=False):
16
+ num_samples = gr.Slider(label='Images',
17
+ minimum=1,
18
+ maximum=max_images,
19
+ value=1,
20
+ step=1)
21
+ image_resolution = gr.Slider(label='Image Resolution',
22
+ minimum=256,
23
+ maximum=768,
24
+ value=512,
25
+ step=256)
26
+ detect_resolution = gr.Slider(label='HED Resolution',
27
+ minimum=128,
28
+ maximum=1024,
29
+ value=512,
30
+ step=1)
31
+ ddim_steps = gr.Slider(label='Steps',
32
+ minimum=1,
33
+ maximum=100,
34
+ value=20,
35
+ step=1)
36
+ scale = gr.Slider(label='Guidance Scale',
37
+ minimum=0.1,
38
+ maximum=30.0,
39
+ value=9.0,
40
+ step=0.1)
41
+ seed = gr.Slider(label='Seed',
42
+ minimum=-1,
43
+ maximum=2147483647,
44
+ step=1,
45
+ randomize=True,
46
+ queue=False)
47
+ eta = gr.Number(label='eta (DDIM)', value=0.0)
48
+ a_prompt = gr.Textbox(
49
+ label='Added Prompt',
50
+ value='best quality, extremely detailed')
51
+ n_prompt = gr.Textbox(
52
+ label='Negative Prompt',
53
+ value=
54
+ 'longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality'
55
+ )
56
+ with gr.Column():
57
+ result_gallery = gr.Gallery(label='Output',
58
+ show_label=False,
59
+ elem_id='gallery').style(
60
+ grid=2, height='auto')
61
+ ips = [
62
+ input_image, prompt, a_prompt, n_prompt, num_samples,
63
+ image_resolution, detect_resolution, ddim_steps, scale, seed, eta
64
+ ]
65
+ run_button.click(fn=process,
66
+ inputs=ips,
67
+ outputs=[result_gallery],
68
+ api_name='hed')
69
+ return demo
gradio_hough2image.py ADDED
@@ -0,0 +1,82 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # This file is adapted from https://github.com/lllyasviel/ControlNet/blob/f4748e3630d8141d7765e2bd9b1e348f47847707/gradio_hough2image.py
2
+ # The original license file is LICENSE.ControlNet in this repo.
3
+ import gradio as gr
4
+
5
+
6
+ def create_demo(process, max_images=12):
7
+ with gr.Blocks() as demo:
8
+ with gr.Row():
9
+ gr.Markdown('## Control Stable Diffusion with Hough Line Maps')
10
+ with gr.Row():
11
+ with gr.Column():
12
+ input_image = gr.Image(source='upload', type='numpy')
13
+ prompt = gr.Textbox(label='Prompt')
14
+ run_button = gr.Button(label='Run')
15
+ with gr.Accordion('Advanced options', open=False):
16
+ num_samples = gr.Slider(label='Images',
17
+ minimum=1,
18
+ maximum=max_images,
19
+ value=1,
20
+ step=1)
21
+ image_resolution = gr.Slider(label='Image Resolution',
22
+ minimum=256,
23
+ maximum=768,
24
+ value=512,
25
+ step=256)
26
+ detect_resolution = gr.Slider(label='Hough Resolution',
27
+ minimum=128,
28
+ maximum=1024,
29
+ value=512,
30
+ step=1)
31
+ value_threshold = gr.Slider(
32
+ label='Hough value threshold (MLSD)',
33
+ minimum=0.01,
34
+ maximum=2.0,
35
+ value=0.1,
36
+ step=0.01)
37
+ distance_threshold = gr.Slider(
38
+ label='Hough distance threshold (MLSD)',
39
+ minimum=0.01,
40
+ maximum=20.0,
41
+ value=0.1,
42
+ step=0.01)
43
+ ddim_steps = gr.Slider(label='Steps',
44
+ minimum=1,
45
+ maximum=100,
46
+ value=20,
47
+ step=1)
48
+ scale = gr.Slider(label='Guidance Scale',
49
+ minimum=0.1,
50
+ maximum=30.0,
51
+ value=9.0,
52
+ step=0.1)
53
+ seed = gr.Slider(label='Seed',
54
+ minimum=-1,
55
+ maximum=2147483647,
56
+ step=1,
57
+ randomize=True,
58
+ queue=False)
59
+ eta = gr.Number(label='eta (DDIM)', value=0.0)
60
+ a_prompt = gr.Textbox(
61
+ label='Added Prompt',
62
+ value='best quality, extremely detailed')
63
+ n_prompt = gr.Textbox(
64
+ label='Negative Prompt',
65
+ value=
66
+ 'longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality'
67
+ )
68
+ with gr.Column():
69
+ result_gallery = gr.Gallery(label='Output',
70
+ show_label=False,
71
+ elem_id='gallery').style(
72
+ grid=2, height='auto')
73
+ ips = [
74
+ input_image, prompt, a_prompt, n_prompt, num_samples,
75
+ image_resolution, detect_resolution, ddim_steps, scale, seed, eta,
76
+ value_threshold, distance_threshold
77
+ ]
78
+ run_button.click(fn=process,
79
+ inputs=ips,
80
+ outputs=[result_gallery],
81
+ api_name='hough')
82
+ return demo
gradio_normal2image.py ADDED
@@ -0,0 +1,76 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # This file is adapted from https://github.com/lllyasviel/ControlNet/blob/f4748e3630d8141d7765e2bd9b1e348f47847707/gradio_normal2image.py
2
+ # The original license file is LICENSE.ControlNet in this repo.
3
+ import gradio as gr
4
+
5
+
6
+ def create_demo(process, max_images=12):
7
+ with gr.Blocks() as demo:
8
+ with gr.Row():
9
+ gr.Markdown('## Control Stable Diffusion with Normal Maps')
10
+ with gr.Row():
11
+ with gr.Column():
12
+ input_image = gr.Image(source='upload', type='numpy')
13
+ prompt = gr.Textbox(label='Prompt')
14
+ run_button = gr.Button(label='Run')
15
+ with gr.Accordion('Advanced options', open=False):
16
+ num_samples = gr.Slider(label='Images',
17
+ minimum=1,
18
+ maximum=max_images,
19
+ value=1,
20
+ step=1)
21
+ image_resolution = gr.Slider(label='Image Resolution',
22
+ minimum=256,
23
+ maximum=768,
24
+ value=512,
25
+ step=256)
26
+ detect_resolution = gr.Slider(label='Normal Resolution',
27
+ minimum=128,
28
+ maximum=1024,
29
+ value=384,
30
+ step=1)
31
+ bg_threshold = gr.Slider(
32
+ label='Normal background threshold',
33
+ minimum=0.0,
34
+ maximum=1.0,
35
+ value=0.4,
36
+ step=0.01)
37
+ ddim_steps = gr.Slider(label='Steps',
38
+ minimum=1,
39
+ maximum=100,
40
+ value=20,
41
+ step=1)
42
+ scale = gr.Slider(label='Guidance Scale',
43
+ minimum=0.1,
44
+ maximum=30.0,
45
+ value=9.0,
46
+ step=0.1)
47
+ seed = gr.Slider(label='Seed',
48
+ minimum=-1,
49
+ maximum=2147483647,
50
+ step=1,
51
+ randomize=True,
52
+ queue=False)
53
+ eta = gr.Number(label='eta (DDIM)', value=0.0)
54
+ a_prompt = gr.Textbox(
55
+ label='Added Prompt',
56
+ value='best quality, extremely detailed')
57
+ n_prompt = gr.Textbox(
58
+ label='Negative Prompt',
59
+ value=
60
+ 'longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality'
61
+ )
62
+ with gr.Column():
63
+ result_gallery = gr.Gallery(label='Output',
64
+ show_label=False,
65
+ elem_id='gallery').style(
66
+ grid=2, height='auto')
67
+ ips = [
68
+ input_image, prompt, a_prompt, n_prompt, num_samples,
69
+ image_resolution, detect_resolution, ddim_steps, scale, seed, eta,
70
+ bg_threshold
71
+ ]
72
+ run_button.click(fn=process,
73
+ inputs=ips,
74
+ outputs=[result_gallery],
75
+ api_name='normal')
76
+ return demo
gradio_pose2image.py ADDED
@@ -0,0 +1,69 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # This file is adapted from https://github.com/lllyasviel/ControlNet/blob/f4748e3630d8141d7765e2bd9b1e348f47847707/gradio_pose2image.py
2
+ # The original license file is LICENSE.ControlNet in this repo.
3
+ import gradio as gr
4
+
5
+
6
+ def create_demo(process, max_images=12):
7
+ with gr.Blocks() as demo:
8
+ with gr.Row():
9
+ gr.Markdown('## Control Stable Diffusion with Human Pose')
10
+ with gr.Row():
11
+ with gr.Column():
12
+ input_image = gr.Image(source='upload', type='numpy')
13
+ prompt = gr.Textbox(label='Prompt')
14
+ run_button = gr.Button(label='Run')
15
+ with gr.Accordion('Advanced options', open=False):
16
+ num_samples = gr.Slider(label='Images',
17
+ minimum=1,
18
+ maximum=max_images,
19
+ value=1,
20
+ step=1)
21
+ image_resolution = gr.Slider(label='Image Resolution',
22
+ minimum=256,
23
+ maximum=768,
24
+ value=512,
25
+ step=256)
26
+ detect_resolution = gr.Slider(label='OpenPose Resolution',
27
+ minimum=128,
28
+ maximum=1024,
29
+ value=512,
30
+ step=1)
31
+ ddim_steps = gr.Slider(label='Steps',
32
+ minimum=1,
33
+ maximum=100,
34
+ value=20,
35
+ step=1)
36
+ scale = gr.Slider(label='Guidance Scale',
37
+ minimum=0.1,
38
+ maximum=30.0,
39
+ value=9.0,
40
+ step=0.1)
41
+ seed = gr.Slider(label='Seed',
42
+ minimum=-1,
43
+ maximum=2147483647,
44
+ step=1,
45
+ randomize=True,
46
+ queue=False)
47
+ eta = gr.Number(label='eta (DDIM)', value=0.0)
48
+ a_prompt = gr.Textbox(
49
+ label='Added Prompt',
50
+ value='best quality, extremely detailed')
51
+ n_prompt = gr.Textbox(
52
+ label='Negative Prompt',
53
+ value=
54
+ 'longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality'
55
+ )
56
+ with gr.Column():
57
+ result_gallery = gr.Gallery(label='Output',
58
+ show_label=False,
59
+ elem_id='gallery').style(
60
+ grid=2, height='auto')
61
+ ips = [
62
+ input_image, prompt, a_prompt, n_prompt, num_samples,
63
+ image_resolution, detect_resolution, ddim_steps, scale, seed, eta
64
+ ]
65
+ run_button.click(fn=process,
66
+ inputs=ips,
67
+ outputs=[result_gallery],
68
+ api_name='pose')
69
+ return demo
gradio_scribble2image.py ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # This file is adapted from https://github.com/lllyasviel/ControlNet/blob/f4748e3630d8141d7765e2bd9b1e348f47847707/gradio_scribble2image.py
2
+ # The original license file is LICENSE.ControlNet in this repo.
3
+ import gradio as gr
4
+
5
+
6
+ def create_demo(process, max_images=12):
7
+ with gr.Blocks() as demo:
8
+ with gr.Row():
9
+ gr.Markdown('## Control Stable Diffusion with Scribble Maps')
10
+ with gr.Row():
11
+ with gr.Column():
12
+ input_image = gr.Image(source='upload', type='numpy')
13
+ prompt = gr.Textbox(label='Prompt')
14
+ run_button = gr.Button(label='Run')
15
+ with gr.Accordion('Advanced options', open=False):
16
+ num_samples = gr.Slider(label='Images',
17
+ minimum=1,
18
+ maximum=max_images,
19
+ value=1,
20
+ step=1)
21
+ image_resolution = gr.Slider(label='Image Resolution',
22
+ minimum=256,
23
+ maximum=768,
24
+ value=512,
25
+ step=256)
26
+ ddim_steps = gr.Slider(label='Steps',
27
+ minimum=1,
28
+ maximum=100,
29
+ value=20,
30
+ step=1)
31
+ scale = gr.Slider(label='Guidance Scale',
32
+ minimum=0.1,
33
+ maximum=30.0,
34
+ value=9.0,
35
+ step=0.1)
36
+ seed = gr.Slider(label='Seed',
37
+ minimum=-1,
38
+ maximum=2147483647,
39
+ step=1,
40
+ randomize=True,
41
+ queue=False)
42
+ eta = gr.Number(label='eta (DDIM)', value=0.0)
43
+ a_prompt = gr.Textbox(
44
+ label='Added Prompt',
45
+ value='best quality, extremely detailed')
46
+ n_prompt = gr.Textbox(
47
+ label='Negative Prompt',
48
+ value=
49
+ 'longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality'
50
+ )
51
+ with gr.Column():
52
+ result_gallery = gr.Gallery(label='Output',
53
+ show_label=False,
54
+ elem_id='gallery').style(
55
+ grid=2, height='auto')
56
+ ips = [
57
+ input_image, prompt, a_prompt, n_prompt, num_samples,
58
+ image_resolution, ddim_steps, scale, seed, eta
59
+ ]
60
+ run_button.click(fn=process,
61
+ inputs=ips,
62
+ outputs=[result_gallery],
63
+ api_name='scribble')
64
+ return demo
gradio_scribble2image_interactive.py ADDED
@@ -0,0 +1,90 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # This file is adapted from https://github.com/lllyasviel/ControlNet/blob/f4748e3630d8141d7765e2bd9b1e348f47847707/gradio_scribble2image_interactive.py
2
+ # The original license file is LICENSE.ControlNet in this repo.
3
+ import gradio as gr
4
+ import numpy as np
5
+
6
+
7
+ def create_canvas(w, h):
8
+ return np.zeros(shape=(h, w, 3), dtype=np.uint8) + 255
9
+
10
+
11
+ def create_demo(process, max_images=12):
12
+ with gr.Blocks() as demo:
13
+ with gr.Row():
14
+ gr.Markdown(
15
+ '## Control Stable Diffusion with Interactive Scribbles')
16
+ with gr.Row():
17
+ with gr.Column():
18
+ canvas_width = gr.Slider(label='Canvas Width',
19
+ minimum=256,
20
+ maximum=1024,
21
+ value=512,
22
+ step=1)
23
+ canvas_height = gr.Slider(label='Canvas Height',
24
+ minimum=256,
25
+ maximum=1024,
26
+ value=512,
27
+ step=1)
28
+ create_button = gr.Button(label='Start',
29
+ value='Open drawing canvas!')
30
+ input_image = gr.Image(source='upload',
31
+ type='numpy',
32
+ tool='sketch')
33
+ gr.Markdown(
34
+ value=
35
+ 'Do not forget to change your brush width to make it thinner. (Gradio do not allow developers to set brush width so you need to do it manually.) '
36
+ 'Just click on the small pencil icon in the upper right corner of the above block.'
37
+ )
38
+ create_button.click(fn=create_canvas,
39
+ inputs=[canvas_width, canvas_height],
40
+ outputs=[input_image],
41
+ queue=False)
42
+ prompt = gr.Textbox(label='Prompt')
43
+ run_button = gr.Button(label='Run')
44
+ with gr.Accordion('Advanced options', open=False):
45
+ num_samples = gr.Slider(label='Images',
46
+ minimum=1,
47
+ maximum=max_images,
48
+ value=1,
49
+ step=1)
50
+ image_resolution = gr.Slider(label='Image Resolution',
51
+ minimum=256,
52
+ maximum=768,
53
+ value=512,
54
+ step=256)
55
+ ddim_steps = gr.Slider(label='Steps',
56
+ minimum=1,
57
+ maximum=100,
58
+ value=20,
59
+ step=1)
60
+ scale = gr.Slider(label='Guidance Scale',
61
+ minimum=0.1,
62
+ maximum=30.0,
63
+ value=9.0,
64
+ step=0.1)
65
+ seed = gr.Slider(label='Seed',
66
+ minimum=-1,
67
+ maximum=2147483647,
68
+ step=1,
69
+ randomize=True,
70
+ queue=False)
71
+ eta = gr.Number(label='eta (DDIM)', value=0.0)
72
+ a_prompt = gr.Textbox(
73
+ label='Added Prompt',
74
+ value='best quality, extremely detailed')
75
+ n_prompt = gr.Textbox(
76
+ label='Negative Prompt',
77
+ value=
78
+ 'longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality'
79
+ )
80
+ with gr.Column():
81
+ result_gallery = gr.Gallery(label='Output',
82
+ show_label=False,
83
+ elem_id='gallery').style(
84
+ grid=2, height='auto')
85
+ ips = [
86
+ input_image, prompt, a_prompt, n_prompt, num_samples,
87
+ image_resolution, ddim_steps, scale, seed, eta
88
+ ]
89
+ run_button.click(fn=process, inputs=ips, outputs=[result_gallery])
90
+ return demo
gradio_seg2image.py ADDED
@@ -0,0 +1,70 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # This file is adapted from https://github.com/lllyasviel/ControlNet/blob/f4748e3630d8141d7765e2bd9b1e348f47847707/gradio_seg2image.py
2
+ # The original license file is LICENSE.ControlNet in this repo.
3
+ import gradio as gr
4
+
5
+
6
+ def create_demo(process, max_images=12):
7
+ with gr.Blocks() as demo:
8
+ with gr.Row():
9
+ gr.Markdown('## Control Stable Diffusion with Segmentation Maps')
10
+ with gr.Row():
11
+ with gr.Column():
12
+ input_image = gr.Image(source='upload', type='numpy')
13
+ prompt = gr.Textbox(label='Prompt')
14
+ run_button = gr.Button(label='Run')
15
+ with gr.Accordion('Advanced options', open=False):
16
+ num_samples = gr.Slider(label='Images',
17
+ minimum=1,
18
+ maximum=max_images,
19
+ value=1,
20
+ step=1)
21
+ image_resolution = gr.Slider(label='Image Resolution',
22
+ minimum=256,
23
+ maximum=768,
24
+ value=512,
25
+ step=256)
26
+ detect_resolution = gr.Slider(
27
+ label='Segmentation Resolution',
28
+ minimum=128,
29
+ maximum=1024,
30
+ value=512,
31
+ step=1)
32
+ ddim_steps = gr.Slider(label='Steps',
33
+ minimum=1,
34
+ maximum=100,
35
+ value=20,
36
+ step=1)
37
+ scale = gr.Slider(label='Guidance Scale',
38
+ minimum=0.1,
39
+ maximum=30.0,
40
+ value=9.0,
41
+ step=0.1)
42
+ seed = gr.Slider(label='Seed',
43
+ minimum=-1,
44
+ maximum=2147483647,
45
+ step=1,
46
+ randomize=True,
47
+ queue=False)
48
+ eta = gr.Number(label='eta (DDIM)', value=0.0)
49
+ a_prompt = gr.Textbox(
50
+ label='Added Prompt',
51
+ value='best quality, extremely detailed')
52
+ n_prompt = gr.Textbox(
53
+ label='Negative Prompt',
54
+ value=
55
+ 'longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality'
56
+ )
57
+ with gr.Column():
58
+ result_gallery = gr.Gallery(label='Output',
59
+ show_label=False,
60
+ elem_id='gallery').style(
61
+ grid=2, height='auto')
62
+ ips = [
63
+ input_image, prompt, a_prompt, n_prompt, num_samples,
64
+ image_resolution, detect_resolution, ddim_steps, scale, seed, eta
65
+ ]
66
+ run_button.click(fn=process,
67
+ inputs=ips,
68
+ outputs=[result_gallery],
69
+ api_name='seg')
70
+ return demo
model.py ADDED
@@ -0,0 +1,766 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # This file is adapted from gradio_*.py in https://github.com/lllyasviel/ControlNet/tree/f4748e3630d8141d7765e2bd9b1e348f47847707
2
+ # The original license file is LICENSE.ControlNet in this repo.
3
+ from __future__ import annotations
4
+
5
+ import pathlib
6
+ import random
7
+ import shlex
8
+ import subprocess
9
+ import sys
10
+
11
+ import cv2
12
+ import einops
13
+ import numpy as np
14
+ import torch
15
+ from pytorch_lightning import seed_everything
16
+
17
+ sys.path.append('ControlNet')
18
+
19
+ import config
20
+ from annotator.canny import apply_canny
21
+ from annotator.hed import apply_hed, nms
22
+ from annotator.midas import apply_midas
23
+ from annotator.mlsd import apply_mlsd
24
+ from annotator.openpose import apply_openpose
25
+ from annotator.uniformer import apply_uniformer
26
+ from annotator.util import HWC3, resize_image
27
+ from cldm.model import create_model, load_state_dict
28
+ from ldm.models.diffusion.ddim import DDIMSampler
29
+ from share import *
30
+
31
+ ORIGINAL_MODEL_NAMES = {
32
+ 'canny': 'control_sd15_canny.pth',
33
+ 'hough': 'control_sd15_mlsd.pth',
34
+ 'hed': 'control_sd15_hed.pth',
35
+ 'scribble': 'control_sd15_scribble.pth',
36
+ 'pose': 'control_sd15_openpose.pth',
37
+ 'seg': 'control_sd15_seg.pth',
38
+ 'depth': 'control_sd15_depth.pth',
39
+ 'normal': 'control_sd15_normal.pth',
40
+ }
41
+ ORIGINAL_WEIGHT_ROOT = 'https://huggingface.co/lllyasviel/ControlNet/resolve/main/models/'
42
+
43
+ LIGHTWEIGHT_MODEL_NAMES = {
44
+ 'canny': 'control_canny-fp16.safetensors',
45
+ 'hough': 'control_mlsd-fp16.safetensors',
46
+ 'hed': 'control_hed-fp16.safetensors',
47
+ 'scribble': 'control_scribble-fp16.safetensors',
48
+ 'pose': 'control_openpose-fp16.safetensors',
49
+ 'seg': 'control_seg-fp16.safetensors',
50
+ 'depth': 'control_depth-fp16.safetensors',
51
+ 'normal': 'control_normal-fp16.safetensors',
52
+ }
53
+ LIGHTWEIGHT_WEIGHT_ROOT = 'https://huggingface.co/webui/ControlNet-modules-safetensors/resolve/main/'
54
+
55
+
56
+ class Model:
57
+ def __init__(self,
58
+ model_config_path: str = 'ControlNet/models/cldm_v15.yaml',
59
+ model_dir: str = 'models',
60
+ use_lightweight: bool = True):
61
+ self.device = torch.device(
62
+ 'cuda:0' if torch.cuda.is_available() else 'cpu')
63
+ self.model = create_model(model_config_path).to(self.device)
64
+ self.ddim_sampler = DDIMSampler(self.model)
65
+ self.task_name = ''
66
+
67
+ self.model_dir = pathlib.Path(model_dir)
68
+ self.model_dir.mkdir(exist_ok=True, parents=True)
69
+
70
+ self.use_lightweight = use_lightweight
71
+ if use_lightweight:
72
+ self.model_names = LIGHTWEIGHT_MODEL_NAMES
73
+ self.weight_root = LIGHTWEIGHT_WEIGHT_ROOT
74
+ base_model_url = 'https://huggingface.co/runwayml/stable-diffusion-v1-5/resolve/main/v1-5-pruned-emaonly.safetensors'
75
+ self.load_base_model(base_model_url)
76
+ else:
77
+ self.model_names = ORIGINAL_MODEL_NAMES
78
+ self.weight_root = ORIGINAL_WEIGHT_ROOT
79
+
80
+ self.download_models()
81
+
82
+ def download_base_model(self, model_url: str) -> pathlib.Path:
83
+ model_name = model_url.split('/')[-1]
84
+ out_path = self.model_dir / model_name
85
+ if not out_path.exists():
86
+ subprocess.run(shlex.split(f'wget {model_url} -O {out_path}'))
87
+ return out_path
88
+
89
+ def load_base_model(self, model_url: str) -> None:
90
+ model_path = self.download_base_model(model_url)
91
+ self.model.load_state_dict(load_state_dict(model_path,
92
+ location=self.device.type),
93
+ strict=False)
94
+
95
+ def load_weight(self, task_name: str) -> None:
96
+ if task_name == self.task_name:
97
+ return
98
+ weight_path = self.get_weight_path(task_name)
99
+ if not self.use_lightweight:
100
+ self.model.load_state_dict(
101
+ load_state_dict(weight_path, location=self.device))
102
+ else:
103
+ self.model.control_model.load_state_dict(
104
+ load_state_dict(weight_path, location=self.device.type))
105
+ self.task_name = task_name
106
+
107
+ def get_weight_path(self, task_name: str) -> str:
108
+ if 'scribble' in task_name:
109
+ task_name = 'scribble'
110
+ return f'{self.model_dir}/{self.model_names[task_name]}'
111
+
112
+ def download_models(self) -> None:
113
+ self.model_dir.mkdir(exist_ok=True, parents=True)
114
+ for name in self.model_names.values():
115
+ out_path = self.model_dir / name
116
+ if out_path.exists():
117
+ continue
118
+ subprocess.run(
119
+ shlex.split(f'wget {self.weight_root}{name} -O {out_path}'))
120
+
121
+ @torch.inference_mode()
122
+ def process_canny(self, input_image, prompt, a_prompt, n_prompt,
123
+ num_samples, image_resolution, ddim_steps, scale, seed,
124
+ eta, low_threshold, high_threshold):
125
+ self.load_weight('canny')
126
+
127
+ img = resize_image(HWC3(input_image), image_resolution)
128
+ H, W, C = img.shape
129
+
130
+ detected_map = apply_canny(img, low_threshold, high_threshold)
131
+ detected_map = HWC3(detected_map)
132
+
133
+ control = torch.from_numpy(detected_map.copy()).float().cuda() / 255.0
134
+ control = torch.stack([control for _ in range(num_samples)], dim=0)
135
+ control = einops.rearrange(control, 'b h w c -> b c h w').clone()
136
+
137
+ if seed == -1:
138
+ seed = random.randint(0, 65535)
139
+ seed_everything(seed)
140
+
141
+ if config.save_memory:
142
+ self.model.low_vram_shift(is_diffusing=False)
143
+
144
+ cond = {
145
+ 'c_concat': [control],
146
+ 'c_crossattn': [
147
+ self.model.get_learned_conditioning(
148
+ [prompt + ', ' + a_prompt] * num_samples)
149
+ ]
150
+ }
151
+ un_cond = {
152
+ 'c_concat': [control],
153
+ 'c_crossattn':
154
+ [self.model.get_learned_conditioning([n_prompt] * num_samples)]
155
+ }
156
+ shape = (4, H // 8, W // 8)
157
+
158
+ if config.save_memory:
159
+ self.model.low_vram_shift(is_diffusing=True)
160
+
161
+ samples, intermediates = self.ddim_sampler.sample(
162
+ ddim_steps,
163
+ num_samples,
164
+ shape,
165
+ cond,
166
+ verbose=False,
167
+ eta=eta,
168
+ unconditional_guidance_scale=scale,
169
+ unconditional_conditioning=un_cond)
170
+
171
+ if config.save_memory:
172
+ self.model.low_vram_shift(is_diffusing=False)
173
+
174
+ x_samples = self.model.decode_first_stage(samples)
175
+ x_samples = (
176
+ einops.rearrange(x_samples, 'b c h w -> b h w c') * 127.5 +
177
+ 127.5).cpu().numpy().clip(0, 255).astype(np.uint8)
178
+
179
+ results = [x_samples[i] for i in range(num_samples)]
180
+ return [255 - detected_map] + results
181
+
182
+ @torch.inference_mode()
183
+ def process_hough(self, input_image, prompt, a_prompt, n_prompt,
184
+ num_samples, image_resolution, detect_resolution,
185
+ ddim_steps, scale, seed, eta, value_threshold,
186
+ distance_threshold):
187
+ self.load_weight('hough')
188
+
189
+ input_image = HWC3(input_image)
190
+ detected_map = apply_mlsd(resize_image(input_image, detect_resolution),
191
+ value_threshold, distance_threshold)
192
+ detected_map = HWC3(detected_map)
193
+ img = resize_image(input_image, image_resolution)
194
+ H, W, C = img.shape
195
+
196
+ detected_map = cv2.resize(detected_map, (W, H),
197
+ interpolation=cv2.INTER_NEAREST)
198
+
199
+ control = torch.from_numpy(detected_map.copy()).float().cuda() / 255.0
200
+ control = torch.stack([control for _ in range(num_samples)], dim=0)
201
+ control = einops.rearrange(control, 'b h w c -> b c h w').clone()
202
+
203
+ if seed == -1:
204
+ seed = random.randint(0, 65535)
205
+ seed_everything(seed)
206
+
207
+ if config.save_memory:
208
+ self.model.low_vram_shift(is_diffusing=False)
209
+
210
+ cond = {
211
+ 'c_concat': [control],
212
+ 'c_crossattn': [
213
+ self.model.get_learned_conditioning(
214
+ [prompt + ', ' + a_prompt] * num_samples)
215
+ ]
216
+ }
217
+ un_cond = {
218
+ 'c_concat': [control],
219
+ 'c_crossattn':
220
+ [self.model.get_learned_conditioning([n_prompt] * num_samples)]
221
+ }
222
+ shape = (4, H // 8, W // 8)
223
+
224
+ if config.save_memory:
225
+ self.model.low_vram_shift(is_diffusing=True)
226
+
227
+ samples, intermediates = self.ddim_sampler.sample(
228
+ ddim_steps,
229
+ num_samples,
230
+ shape,
231
+ cond,
232
+ verbose=False,
233
+ eta=eta,
234
+ unconditional_guidance_scale=scale,
235
+ unconditional_conditioning=un_cond)
236
+
237
+ if config.save_memory:
238
+ self.model.low_vram_shift(is_diffusing=False)
239
+
240
+ x_samples = self.model.decode_first_stage(samples)
241
+ x_samples = (
242
+ einops.rearrange(x_samples, 'b c h w -> b h w c') * 127.5 +
243
+ 127.5).cpu().numpy().clip(0, 255).astype(np.uint8)
244
+
245
+ results = [x_samples[i] for i in range(num_samples)]
246
+ return [
247
+ 255 - cv2.dilate(detected_map,
248
+ np.ones(shape=(3, 3), dtype=np.uint8),
249
+ iterations=1)
250
+ ] + results
251
+
252
+ @torch.inference_mode()
253
+ def process_hed(self, input_image, prompt, a_prompt, n_prompt, num_samples,
254
+ image_resolution, detect_resolution, ddim_steps, scale,
255
+ seed, eta):
256
+ self.load_weight('hed')
257
+
258
+ input_image = HWC3(input_image)
259
+ detected_map = apply_hed(resize_image(input_image, detect_resolution))
260
+ detected_map = HWC3(detected_map)
261
+ img = resize_image(input_image, image_resolution)
262
+ H, W, C = img.shape
263
+
264
+ detected_map = cv2.resize(detected_map, (W, H),
265
+ interpolation=cv2.INTER_LINEAR)
266
+
267
+ control = torch.from_numpy(detected_map.copy()).float().cuda() / 255.0
268
+ control = torch.stack([control for _ in range(num_samples)], dim=0)
269
+ control = einops.rearrange(control, 'b h w c -> b c h w').clone()
270
+
271
+ if seed == -1:
272
+ seed = random.randint(0, 65535)
273
+ seed_everything(seed)
274
+
275
+ if config.save_memory:
276
+ self.model.low_vram_shift(is_diffusing=False)
277
+
278
+ cond = {
279
+ 'c_concat': [control],
280
+ 'c_crossattn': [
281
+ self.model.get_learned_conditioning(
282
+ [prompt + ', ' + a_prompt] * num_samples)
283
+ ]
284
+ }
285
+ un_cond = {
286
+ 'c_concat': [control],
287
+ 'c_crossattn':
288
+ [self.model.get_learned_conditioning([n_prompt] * num_samples)]
289
+ }
290
+ shape = (4, H // 8, W // 8)
291
+
292
+ if config.save_memory:
293
+ self.model.low_vram_shift(is_diffusing=True)
294
+
295
+ samples, intermediates = self.ddim_sampler.sample(
296
+ ddim_steps,
297
+ num_samples,
298
+ shape,
299
+ cond,
300
+ verbose=False,
301
+ eta=eta,
302
+ unconditional_guidance_scale=scale,
303
+ unconditional_conditioning=un_cond)
304
+
305
+ if config.save_memory:
306
+ self.model.low_vram_shift(is_diffusing=False)
307
+
308
+ x_samples = self.model.decode_first_stage(samples)
309
+ x_samples = (
310
+ einops.rearrange(x_samples, 'b c h w -> b h w c') * 127.5 +
311
+ 127.5).cpu().numpy().clip(0, 255).astype(np.uint8)
312
+
313
+ results = [x_samples[i] for i in range(num_samples)]
314
+ return [detected_map] + results
315
+
316
+ @torch.inference_mode()
317
+ def process_scribble(self, input_image, prompt, a_prompt, n_prompt,
318
+ num_samples, image_resolution, ddim_steps, scale,
319
+ seed, eta):
320
+ self.load_weight('scribble')
321
+
322
+ img = resize_image(HWC3(input_image), image_resolution)
323
+ H, W, C = img.shape
324
+
325
+ detected_map = np.zeros_like(img, dtype=np.uint8)
326
+ detected_map[np.min(img, axis=2) < 127] = 255
327
+
328
+ control = torch.from_numpy(detected_map.copy()).float().cuda() / 255.0
329
+ control = torch.stack([control for _ in range(num_samples)], dim=0)
330
+ control = einops.rearrange(control, 'b h w c -> b c h w').clone()
331
+
332
+ if seed == -1:
333
+ seed = random.randint(0, 65535)
334
+ seed_everything(seed)
335
+
336
+ if config.save_memory:
337
+ self.model.low_vram_shift(is_diffusing=False)
338
+
339
+ cond = {
340
+ 'c_concat': [control],
341
+ 'c_crossattn': [
342
+ self.model.get_learned_conditioning(
343
+ [prompt + ', ' + a_prompt] * num_samples)
344
+ ]
345
+ }
346
+ un_cond = {
347
+ 'c_concat': [control],
348
+ 'c_crossattn':
349
+ [self.model.get_learned_conditioning([n_prompt] * num_samples)]
350
+ }
351
+ shape = (4, H // 8, W // 8)
352
+
353
+ if config.save_memory:
354
+ self.model.low_vram_shift(is_diffusing=True)
355
+
356
+ samples, intermediates = self.ddim_sampler.sample(
357
+ ddim_steps,
358
+ num_samples,
359
+ shape,
360
+ cond,
361
+ verbose=False,
362
+ eta=eta,
363
+ unconditional_guidance_scale=scale,
364
+ unconditional_conditioning=un_cond)
365
+
366
+ if config.save_memory:
367
+ self.model.low_vram_shift(is_diffusing=False)
368
+
369
+ x_samples = self.model.decode_first_stage(samples)
370
+ x_samples = (
371
+ einops.rearrange(x_samples, 'b c h w -> b h w c') * 127.5 +
372
+ 127.5).cpu().numpy().clip(0, 255).astype(np.uint8)
373
+
374
+ results = [x_samples[i] for i in range(num_samples)]
375
+ return [255 - detected_map] + results
376
+
377
+ @torch.inference_mode()
378
+ def process_scribble_interactive(self, input_image, prompt, a_prompt,
379
+ n_prompt, num_samples, image_resolution,
380
+ ddim_steps, scale, seed, eta):
381
+ self.load_weight('scribble')
382
+
383
+ img = resize_image(HWC3(input_image['mask'][:, :, 0]),
384
+ image_resolution)
385
+ H, W, C = img.shape
386
+
387
+ detected_map = np.zeros_like(img, dtype=np.uint8)
388
+ detected_map[np.min(img, axis=2) > 127] = 255
389
+
390
+ control = torch.from_numpy(detected_map.copy()).float().cuda() / 255.0
391
+ control = torch.stack([control for _ in range(num_samples)], dim=0)
392
+ control = einops.rearrange(control, 'b h w c -> b c h w').clone()
393
+
394
+ if seed == -1:
395
+ seed = random.randint(0, 65535)
396
+ seed_everything(seed)
397
+
398
+ if config.save_memory:
399
+ self.model.low_vram_shift(is_diffusing=False)
400
+
401
+ cond = {
402
+ 'c_concat': [control],
403
+ 'c_crossattn': [
404
+ self.model.get_learned_conditioning(
405
+ [prompt + ', ' + a_prompt] * num_samples)
406
+ ]
407
+ }
408
+ un_cond = {
409
+ 'c_concat': [control],
410
+ 'c_crossattn':
411
+ [self.model.get_learned_conditioning([n_prompt] * num_samples)]
412
+ }
413
+ shape = (4, H // 8, W // 8)
414
+
415
+ if config.save_memory:
416
+ self.model.low_vram_shift(is_diffusing=True)
417
+
418
+ samples, intermediates = self.ddim_sampler.sample(
419
+ ddim_steps,
420
+ num_samples,
421
+ shape,
422
+ cond,
423
+ verbose=False,
424
+ eta=eta,
425
+ unconditional_guidance_scale=scale,
426
+ unconditional_conditioning=un_cond)
427
+
428
+ if config.save_memory:
429
+ self.model.low_vram_shift(is_diffusing=False)
430
+
431
+ x_samples = self.model.decode_first_stage(samples)
432
+ x_samples = (
433
+ einops.rearrange(x_samples, 'b c h w -> b h w c') * 127.5 +
434
+ 127.5).cpu().numpy().clip(0, 255).astype(np.uint8)
435
+
436
+ results = [x_samples[i] for i in range(num_samples)]
437
+ return [255 - detected_map] + results
438
+
439
+ @torch.inference_mode()
440
+ def process_fake_scribble(self, input_image, prompt, a_prompt, n_prompt,
441
+ num_samples, image_resolution, detect_resolution,
442
+ ddim_steps, scale, seed, eta):
443
+ self.load_weight('scribble')
444
+
445
+ input_image = HWC3(input_image)
446
+ detected_map = apply_hed(resize_image(input_image, detect_resolution))
447
+ detected_map = HWC3(detected_map)
448
+ img = resize_image(input_image, image_resolution)
449
+ H, W, C = img.shape
450
+
451
+ detected_map = cv2.resize(detected_map, (W, H),
452
+ interpolation=cv2.INTER_LINEAR)
453
+ detected_map = nms(detected_map, 127, 3.0)
454
+ detected_map = cv2.GaussianBlur(detected_map, (0, 0), 3.0)
455
+ detected_map[detected_map > 4] = 255
456
+ detected_map[detected_map < 255] = 0
457
+
458
+ control = torch.from_numpy(detected_map.copy()).float().cuda() / 255.0
459
+ control = torch.stack([control for _ in range(num_samples)], dim=0)
460
+ control = einops.rearrange(control, 'b h w c -> b c h w').clone()
461
+
462
+ if seed == -1:
463
+ seed = random.randint(0, 65535)
464
+ seed_everything(seed)
465
+
466
+ if config.save_memory:
467
+ self.model.low_vram_shift(is_diffusing=False)
468
+
469
+ cond = {
470
+ 'c_concat': [control],
471
+ 'c_crossattn': [
472
+ self.model.get_learned_conditioning(
473
+ [prompt + ', ' + a_prompt] * num_samples)
474
+ ]
475
+ }
476
+ un_cond = {
477
+ 'c_concat': [control],
478
+ 'c_crossattn':
479
+ [self.model.get_learned_conditioning([n_prompt] * num_samples)]
480
+ }
481
+ shape = (4, H // 8, W // 8)
482
+
483
+ if config.save_memory:
484
+ self.model.low_vram_shift(is_diffusing=True)
485
+
486
+ samples, intermediates = self.ddim_sampler.sample(
487
+ ddim_steps,
488
+ num_samples,
489
+ shape,
490
+ cond,
491
+ verbose=False,
492
+ eta=eta,
493
+ unconditional_guidance_scale=scale,
494
+ unconditional_conditioning=un_cond)
495
+
496
+ if config.save_memory:
497
+ self.model.low_vram_shift(is_diffusing=False)
498
+
499
+ x_samples = self.model.decode_first_stage(samples)
500
+ x_samples = (
501
+ einops.rearrange(x_samples, 'b c h w -> b h w c') * 127.5 +
502
+ 127.5).cpu().numpy().clip(0, 255).astype(np.uint8)
503
+
504
+ results = [x_samples[i] for i in range(num_samples)]
505
+ return [255 - detected_map] + results
506
+
507
+ @torch.inference_mode()
508
+ def process_pose(self, input_image, prompt, a_prompt, n_prompt,
509
+ num_samples, image_resolution, detect_resolution,
510
+ ddim_steps, scale, seed, eta):
511
+ self.load_weight('pose')
512
+
513
+ input_image = HWC3(input_image)
514
+ detected_map, _ = apply_openpose(
515
+ resize_image(input_image, detect_resolution))
516
+ detected_map = HWC3(detected_map)
517
+ img = resize_image(input_image, image_resolution)
518
+ H, W, C = img.shape
519
+
520
+ detected_map = cv2.resize(detected_map, (W, H),
521
+ interpolation=cv2.INTER_NEAREST)
522
+
523
+ control = torch.from_numpy(detected_map.copy()).float().cuda() / 255.0
524
+ control = torch.stack([control for _ in range(num_samples)], dim=0)
525
+ control = einops.rearrange(control, 'b h w c -> b c h w').clone()
526
+
527
+ if seed == -1:
528
+ seed = random.randint(0, 65535)
529
+ seed_everything(seed)
530
+
531
+ if config.save_memory:
532
+ self.model.low_vram_shift(is_diffusing=False)
533
+
534
+ cond = {
535
+ 'c_concat': [control],
536
+ 'c_crossattn': [
537
+ self.model.get_learned_conditioning(
538
+ [prompt + ', ' + a_prompt] * num_samples)
539
+ ]
540
+ }
541
+ un_cond = {
542
+ 'c_concat': [control],
543
+ 'c_crossattn':
544
+ [self.model.get_learned_conditioning([n_prompt] * num_samples)]
545
+ }
546
+ shape = (4, H // 8, W // 8)
547
+
548
+ if config.save_memory:
549
+ self.model.low_vram_shift(is_diffusing=True)
550
+
551
+ samples, intermediates = self.ddim_sampler.sample(
552
+ ddim_steps,
553
+ num_samples,
554
+ shape,
555
+ cond,
556
+ verbose=False,
557
+ eta=eta,
558
+ unconditional_guidance_scale=scale,
559
+ unconditional_conditioning=un_cond)
560
+
561
+ if config.save_memory:
562
+ self.model.low_vram_shift(is_diffusing=False)
563
+
564
+ x_samples = self.model.decode_first_stage(samples)
565
+ x_samples = (
566
+ einops.rearrange(x_samples, 'b c h w -> b h w c') * 127.5 +
567
+ 127.5).cpu().numpy().clip(0, 255).astype(np.uint8)
568
+
569
+ results = [x_samples[i] for i in range(num_samples)]
570
+ return [detected_map] + results
571
+
572
+ @torch.inference_mode()
573
+ def process_seg(self, input_image, prompt, a_prompt, n_prompt, num_samples,
574
+ image_resolution, detect_resolution, ddim_steps, scale,
575
+ seed, eta):
576
+ self.load_weight('seg')
577
+
578
+ input_image = HWC3(input_image)
579
+ detected_map = apply_uniformer(
580
+ resize_image(input_image, detect_resolution))
581
+ img = resize_image(input_image, image_resolution)
582
+ H, W, C = img.shape
583
+
584
+ detected_map = cv2.resize(detected_map, (W, H),
585
+ interpolation=cv2.INTER_NEAREST)
586
+
587
+ control = torch.from_numpy(detected_map.copy()).float().cuda() / 255.0
588
+ control = torch.stack([control for _ in range(num_samples)], dim=0)
589
+ control = einops.rearrange(control, 'b h w c -> b c h w').clone()
590
+
591
+ if seed == -1:
592
+ seed = random.randint(0, 65535)
593
+ seed_everything(seed)
594
+
595
+ if config.save_memory:
596
+ self.model.low_vram_shift(is_diffusing=False)
597
+
598
+ cond = {
599
+ 'c_concat': [control],
600
+ 'c_crossattn': [
601
+ self.model.get_learned_conditioning(
602
+ [prompt + ', ' + a_prompt] * num_samples)
603
+ ]
604
+ }
605
+ un_cond = {
606
+ 'c_concat': [control],
607
+ 'c_crossattn':
608
+ [self.model.get_learned_conditioning([n_prompt] * num_samples)]
609
+ }
610
+ shape = (4, H // 8, W // 8)
611
+
612
+ if config.save_memory:
613
+ self.model.low_vram_shift(is_diffusing=True)
614
+
615
+ samples, intermediates = self.ddim_sampler.sample(
616
+ ddim_steps,
617
+ num_samples,
618
+ shape,
619
+ cond,
620
+ verbose=False,
621
+ eta=eta,
622
+ unconditional_guidance_scale=scale,
623
+ unconditional_conditioning=un_cond)
624
+
625
+ if config.save_memory:
626
+ self.model.low_vram_shift(is_diffusing=False)
627
+
628
+ x_samples = self.model.decode_first_stage(samples)
629
+ x_samples = (
630
+ einops.rearrange(x_samples, 'b c h w -> b h w c') * 127.5 +
631
+ 127.5).cpu().numpy().clip(0, 255).astype(np.uint8)
632
+
633
+ results = [x_samples[i] for i in range(num_samples)]
634
+ return [detected_map] + results
635
+
636
+ @torch.inference_mode()
637
+ def process_depth(self, input_image, prompt, a_prompt, n_prompt,
638
+ num_samples, image_resolution, detect_resolution,
639
+ ddim_steps, scale, seed, eta):
640
+ self.load_weight('depth')
641
+
642
+ input_image = HWC3(input_image)
643
+ detected_map, _ = apply_midas(
644
+ resize_image(input_image, detect_resolution))
645
+ detected_map = HWC3(detected_map)
646
+ img = resize_image(input_image, image_resolution)
647
+ H, W, C = img.shape
648
+
649
+ detected_map = cv2.resize(detected_map, (W, H),
650
+ interpolation=cv2.INTER_LINEAR)
651
+
652
+ control = torch.from_numpy(detected_map.copy()).float().cuda() / 255.0
653
+ control = torch.stack([control for _ in range(num_samples)], dim=0)
654
+ control = einops.rearrange(control, 'b h w c -> b c h w').clone()
655
+
656
+ if seed == -1:
657
+ seed = random.randint(0, 65535)
658
+ seed_everything(seed)
659
+
660
+ if config.save_memory:
661
+ self.model.low_vram_shift(is_diffusing=False)
662
+
663
+ cond = {
664
+ 'c_concat': [control],
665
+ 'c_crossattn': [
666
+ self.model.get_learned_conditioning(
667
+ [prompt + ', ' + a_prompt] * num_samples)
668
+ ]
669
+ }
670
+ un_cond = {
671
+ 'c_concat': [control],
672
+ 'c_crossattn':
673
+ [self.model.get_learned_conditioning([n_prompt] * num_samples)]
674
+ }
675
+ shape = (4, H // 8, W // 8)
676
+
677
+ if config.save_memory:
678
+ self.model.low_vram_shift(is_diffusing=True)
679
+
680
+ samples, intermediates = self.ddim_sampler.sample(
681
+ ddim_steps,
682
+ num_samples,
683
+ shape,
684
+ cond,
685
+ verbose=False,
686
+ eta=eta,
687
+ unconditional_guidance_scale=scale,
688
+ unconditional_conditioning=un_cond)
689
+
690
+ if config.save_memory:
691
+ self.model.low_vram_shift(is_diffusing=False)
692
+
693
+ x_samples = self.model.decode_first_stage(samples)
694
+ x_samples = (
695
+ einops.rearrange(x_samples, 'b c h w -> b h w c') * 127.5 +
696
+ 127.5).cpu().numpy().clip(0, 255).astype(np.uint8)
697
+
698
+ results = [x_samples[i] for i in range(num_samples)]
699
+ return [detected_map] + results
700
+
701
+ @torch.inference_mode()
702
+ def process_normal(self, input_image, prompt, a_prompt, n_prompt,
703
+ num_samples, image_resolution, detect_resolution,
704
+ ddim_steps, scale, seed, eta, bg_threshold):
705
+ self.load_weight('normal')
706
+
707
+ input_image = HWC3(input_image)
708
+ _, detected_map = apply_midas(resize_image(input_image,
709
+ detect_resolution),
710
+ bg_th=bg_threshold)
711
+ detected_map = HWC3(detected_map)
712
+ img = resize_image(input_image, image_resolution)
713
+ H, W, C = img.shape
714
+
715
+ detected_map = cv2.resize(detected_map, (W, H),
716
+ interpolation=cv2.INTER_LINEAR)
717
+
718
+ control = torch.from_numpy(
719
+ detected_map[:, :, ::-1].copy()).float().cuda() / 255.0
720
+ control = torch.stack([control for _ in range(num_samples)], dim=0)
721
+ control = einops.rearrange(control, 'b h w c -> b c h w').clone()
722
+
723
+ if seed == -1:
724
+ seed = random.randint(0, 65535)
725
+ seed_everything(seed)
726
+
727
+ if config.save_memory:
728
+ self.model.low_vram_shift(is_diffusing=False)
729
+
730
+ cond = {
731
+ 'c_concat': [control],
732
+ 'c_crossattn': [
733
+ self.model.get_learned_conditioning(
734
+ [prompt + ', ' + a_prompt] * num_samples)
735
+ ]
736
+ }
737
+ un_cond = {
738
+ 'c_concat': [control],
739
+ 'c_crossattn':
740
+ [self.model.get_learned_conditioning([n_prompt] * num_samples)]
741
+ }
742
+ shape = (4, H // 8, W // 8)
743
+
744
+ if config.save_memory:
745
+ self.model.low_vram_shift(is_diffusing=True)
746
+
747
+ samples, intermediates = self.ddim_sampler.sample(
748
+ ddim_steps,
749
+ num_samples,
750
+ shape,
751
+ cond,
752
+ verbose=False,
753
+ eta=eta,
754
+ unconditional_guidance_scale=scale,
755
+ unconditional_conditioning=un_cond)
756
+
757
+ if config.save_memory:
758
+ self.model.low_vram_shift(is_diffusing=False)
759
+
760
+ x_samples = self.model.decode_first_stage(samples)
761
+ x_samples = (
762
+ einops.rearrange(x_samples, 'b c h w -> b h w c') * 127.5 +
763
+ 127.5).cpu().numpy().clip(0, 255).astype(np.uint8)
764
+
765
+ results = [x_samples[i] for i in range(num_samples)]
766
+ return [detected_map] + results
patch ADDED
@@ -0,0 +1,115 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ diff --git a/annotator/hed/__init__.py b/annotator/hed/__init__.py
2
+ index 42d8dc6..1587035 100644
3
+ --- a/annotator/hed/__init__.py
4
+ +++ b/annotator/hed/__init__.py
5
+ @@ -1,8 +1,12 @@
6
+ +import pathlib
7
+ +
8
+ import numpy as np
9
+ import cv2
10
+ import torch
11
+ from einops import rearrange
12
+
13
+ +root_dir = pathlib.Path(__file__).parents[2]
14
+ +
15
+
16
+ class Network(torch.nn.Module):
17
+ def __init__(self):
18
+ @@ -64,7 +68,7 @@ class Network(torch.nn.Module):
19
+ torch.nn.Sigmoid()
20
+ )
21
+
22
+ - self.load_state_dict({strKey.replace('module', 'net'): tenWeight for strKey, tenWeight in torch.load('./annotator/ckpts/network-bsds500.pth').items()})
23
+ + self.load_state_dict({strKey.replace('module', 'net'): tenWeight for strKey, tenWeight in torch.load(f'{root_dir}/annotator/ckpts/network-bsds500.pth').items()})
24
+ # end
25
+
26
+ def forward(self, tenInput):
27
+ diff --git a/annotator/midas/api.py b/annotator/midas/api.py
28
+ index 9fa305e..d8594ea 100644
29
+ --- a/annotator/midas/api.py
30
+ +++ b/annotator/midas/api.py
31
+ @@ -1,5 +1,7 @@
32
+ # based on https://github.com/isl-org/MiDaS
33
+
34
+ +import pathlib
35
+ +
36
+ import cv2
37
+ import torch
38
+ import torch.nn as nn
39
+ @@ -10,10 +12,11 @@ from .midas.midas_net import MidasNet
40
+ from .midas.midas_net_custom import MidasNet_small
41
+ from .midas.transforms import Resize, NormalizeImage, PrepareForNet
42
+
43
+ +root_dir = pathlib.Path(__file__).parents[2]
44
+
45
+ ISL_PATHS = {
46
+ - "dpt_large": "annotator/ckpts/dpt_large-midas-2f21e586.pt",
47
+ - "dpt_hybrid": "annotator/ckpts/dpt_hybrid-midas-501f0c75.pt",
48
+ + "dpt_large": f"{root_dir}/annotator/ckpts/dpt_large-midas-2f21e586.pt",
49
+ + "dpt_hybrid": f"{root_dir}/annotator/ckpts/dpt_hybrid-midas-501f0c75.pt",
50
+ "midas_v21": "",
51
+ "midas_v21_small": "",
52
+ }
53
+ diff --git a/annotator/mlsd/__init__.py b/annotator/mlsd/__init__.py
54
+ index 75db717..f310fe6 100644
55
+ --- a/annotator/mlsd/__init__.py
56
+ +++ b/annotator/mlsd/__init__.py
57
+ @@ -1,3 +1,5 @@
58
+ +import pathlib
59
+ +
60
+ import cv2
61
+ import numpy as np
62
+ import torch
63
+ @@ -8,8 +10,9 @@ from .models.mbv2_mlsd_tiny import MobileV2_MLSD_Tiny
64
+ from .models.mbv2_mlsd_large import MobileV2_MLSD_Large
65
+ from .utils import pred_lines
66
+
67
+ +root_dir = pathlib.Path(__file__).parents[2]
68
+
69
+ -model_path = './annotator/ckpts/mlsd_large_512_fp32.pth'
70
+ +model_path = f'{root_dir}/annotator/ckpts/mlsd_large_512_fp32.pth'
71
+ model = MobileV2_MLSD_Large()
72
+ model.load_state_dict(torch.load(model_path), strict=True)
73
+ model = model.cuda().eval()
74
+ diff --git a/annotator/openpose/__init__.py b/annotator/openpose/__init__.py
75
+ index 47d50a5..2369eed 100644
76
+ --- a/annotator/openpose/__init__.py
77
+ +++ b/annotator/openpose/__init__.py
78
+ @@ -1,4 +1,5 @@
79
+ import os
80
+ +import pathlib
81
+ os.environ["KMP_DUPLICATE_LIB_OK"]="TRUE"
82
+
83
+ import torch
84
+ @@ -7,8 +8,10 @@ from . import util
85
+ from .body import Body
86
+ from .hand import Hand
87
+
88
+ -body_estimation = Body('./annotator/ckpts/body_pose_model.pth')
89
+ -hand_estimation = Hand('./annotator/ckpts/hand_pose_model.pth')
90
+ +root_dir = pathlib.Path(__file__).parents[2]
91
+ +
92
+ +body_estimation = Body(f'{root_dir}/annotator/ckpts/body_pose_model.pth')
93
+ +hand_estimation = Hand(f'{root_dir}/annotator/ckpts/hand_pose_model.pth')
94
+
95
+
96
+ def apply_openpose(oriImg, hand=False):
97
+ diff --git a/annotator/uniformer/__init__.py b/annotator/uniformer/__init__.py
98
+ index 500e53c..4061dbe 100644
99
+ --- a/annotator/uniformer/__init__.py
100
+ +++ b/annotator/uniformer/__init__.py
101
+ @@ -1,9 +1,12 @@
102
+ +import pathlib
103
+ +
104
+ from annotator.uniformer.mmseg.apis import init_segmentor, inference_segmentor, show_result_pyplot
105
+ from annotator.uniformer.mmseg.core.evaluation import get_palette
106
+
107
+ +root_dir = pathlib.Path(__file__).parents[2]
108
+
109
+ -checkpoint_file = "annotator/ckpts/upernet_global_small.pth"
110
+ -config_file = 'annotator/uniformer/exp/upernet_global_small/config.py'
111
+ +checkpoint_file = f"{root_dir}/annotator/ckpts/upernet_global_small.pth"
112
+ +config_file = f'{root_dir}/annotator/uniformer/exp/upernet_global_small/config.py'
113
+ model = init_segmentor(config_file, checkpoint_file).cuda()
114
+
115
+
requirements.txt ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ addict==2.4.0
2
+ albumentations==1.3.0
3
+ einops==0.6.0
4
+ gradio==3.18.0
5
+ imageio==2.25.0
6
+ imageio-ffmpeg==0.4.8
7
+ kornia==0.6.9
8
+ omegaconf==2.3.0
9
+ open-clip-torch==2.13.0
10
+ opencv-contrib-python==4.7.0.68
11
+ opencv-python-headless==4.7.0.68
12
+ prettytable==3.6.0
13
+ pytorch-lightning==1.9.0
14
+ safetensors==0.2.8
15
+ timm==0.6.12
16
+ torch==1.13.1
17
+ torchvision==0.14.1
18
+ transformers==4.26.1
19
+ xformers==0.0.16
20
+ yapf==0.32.0
style.css ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ h1 {
2
+ text-align: center;
3
+ }