rttrsabc zRzRzRzRzRzRzR commited on
Commit
edc500f
0 Parent(s):

Duplicate from THUDM/CogVideoX-2b

Browse files

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

.gitattributes 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
.gitignore ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ output/
2
+ *__pycache__/
3
+ samples*/
4
+ runs/
5
+ checkpoints/
6
+ master_ip
7
+ logs/
8
+ *.DS_Store
9
+ .idea
LICENSE 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 2024 CogVideo Model Team @ Zhipu AI
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 +1,332 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ language:
4
+ - en
5
+ tags:
6
+ - cogvideox
7
+ - video-generation
8
+ - thudm
9
+ - text-to-video
10
+ inference: false
11
+ ---
12
+
13
+ # CogVideoX-2B
14
+
15
+ <p style="text-align: center;">
16
+ <div align="center">
17
+ <img src=https://github.com/THUDM/CogVideo/raw/main/resources/logo.svg width="50%"/>
18
+ </div>
19
+ <p align="center">
20
+ <a href="https://huggingface.co/THUDM/CogVideoX-2b/blob/main/README_zh.md">📄 中文阅读</a> |
21
+ <a href="https://huggingface.co/spaces/THUDM/CogVideoX-2B-Space">🤗 Huggingface Space</a> |
22
+ <a href="https://github.com/THUDM/CogVideo">🌐 Github </a> |
23
+ <a href="https://arxiv.org/pdf/2408.06072">📜 arxiv </a>
24
+ </p>
25
+ <p align="center">
26
+ 📍 Visit <a href="https://chatglm.cn/video?lang=en?fr=osm_cogvideo">QingYing</a> and <a href="https://open.bigmodel.cn/?utm_campaign=open&_channel_track_key=OWTVNma9">API Platform</a> to experience commercial video generation models.
27
+ </p>
28
+
29
+ ## Demo Show
30
+
31
+ <!DOCTYPE html>
32
+ <html lang="en">
33
+ <head>
34
+ <meta charset="UTF-8">
35
+ <meta name="viewport" content="width=device-width, initial-scale=1.0">
36
+ <title>Video Gallery with Captions</title>
37
+ <style>
38
+ .video-container {
39
+ display: flex;
40
+ flex-wrap: wrap;
41
+ justify-content: space-around;
42
+ }
43
+ .video-item {
44
+ width: 45%;
45
+ margin-bottom: 20px;
46
+ transition: transform 0.3s;
47
+ }
48
+ .video-item:hover {
49
+ transform: scale(1.1);
50
+ }
51
+ .caption {
52
+ text-align: center;
53
+ margin-top: 10px;
54
+ font-size: 11px;
55
+ }
56
+ </style>
57
+ </head>
58
+ <body>
59
+ <div class="video-container">
60
+ <div class="video-item">
61
+ <video width="100%" controls>
62
+ <source src="https://github.com/THUDM/CogVideo/raw/main/resources/videos/1.mp4" type="video/mp4">
63
+ </video>
64
+ <div class="caption">A detailed wooden toy ship with intricately carved masts and sails is seen gliding smoothly over a plush, blue carpet that mimics the waves of the sea. The ship's hull is painted a rich brown, with tiny windows. The carpet, soft and textured, provides a perfect backdrop, resembling an oceanic expanse. Surrounding the ship are various other toys and children's items, hinting at a playful environment. The scene captures the innocence and imagination of childhood, with the toy ship's journey symbolizing endless adventures in a whimsical, indoor setting.</div>
65
+ </div>
66
+ <div class="video-item">
67
+ <video width="100%" controls>
68
+ <source src="https://github.com/THUDM/CogVideo/raw/main/resources/videos/2.mp4" type="video/mp4">
69
+ </video>
70
+ <div class="caption">The camera follows behind a white vintage SUV with a black roof rack as it speeds up a steep dirt road surrounded by pine trees on a steep mountain slope, dust kicks up from it’s tires, the sunlight shines on the SUV as it speeds along the dirt road, casting a warm glow over the scene. The dirt road curves gently into the distance, with no other cars or vehicles in sight. The trees on either side of the road are redwoods, with patches of greenery scattered throughout. The car is seen from the rear following the curve with ease, making it seem as if it is on a rugged drive through the rugged terrain. The dirt road itself is surrounded by steep hills and mountains, with a clear blue sky above with wispy clouds.</div>
71
+ </div>
72
+ <div class="video-item">
73
+ <video width="100%" controls>
74
+ <source src="https://github.com/THUDM/CogVideo/raw/main/resources/videos/3.mp4" type="video/mp4">
75
+ </video>
76
+ <div class="caption">A street artist, clad in a worn-out denim jacket and a colorful bandana, stands before a vast concrete wall in the heart, holding a can of spray paint, spray-painting a colorful bird on a mottled wall.</div>
77
+ </div>
78
+ <div class="video-item">
79
+ <video width="100%" controls>
80
+ <source src="https://github.com/THUDM/CogVideo/raw/main/resources/videos/4.mp4" type="video/mp4">
81
+ </video>
82
+ <div class="caption"> In the haunting backdrop of a war-torn city, where ruins and crumbled walls tell a story of devastation, a poignant close-up frames a young girl. Her face is smudged with ash, a silent testament to the chaos around her. Her eyes glistening with a mix of sorrow and resilience, capturing the raw emotion of a world that has lost its innocence to the ravages of conflict.</div>
83
+ </div>
84
+ </div>
85
+ </body>
86
+ </html>
87
+
88
+ ## Model Introduction
89
+
90
+ CogVideoX is an open-source version of the video generation model originating
91
+ from [QingYing](https://chatglm.cn/video?lang=en?fr=osm_cogvideo). The table below displays the list of video generation
92
+ models we currently offer, along with their foundational information.
93
+
94
+ <table style="border-collapse: collapse; width: 100%;">
95
+ <tr>
96
+ <th style="text-align: center;">Model Name</th>
97
+ <th style="text-align: center;">CogVideoX-2B (This Repository)</th>
98
+ <th style="text-align: center;">CogVideoX-5B</th>
99
+ </tr>
100
+ <tr>
101
+ <td style="text-align: center;">Model Description</td>
102
+ <td style="text-align: center;">Entry-level model, balancing compatibility. Low cost for running and secondary development.</td>
103
+ <td style="text-align: center;">Larger model with higher video generation quality and better visual effects.</td>
104
+ </tr>
105
+ <tr>
106
+ <td style="text-align: center;">Inference Precision</td>
107
+ <td style="text-align: center;"><b>FP16* (Recommended)</b>, BF16, FP32, FP8*, INT8, no support for INT4</td>
108
+ <td style="text-align: center;"><b>BF16 (Recommended)</b>, FP16, FP32, FP8*, INT8, no support for INT4</td>
109
+ </tr>
110
+ <tr>
111
+ <td style="text-align: center;">Single GPU VRAM Consumption<br></td>
112
+ <td style="text-align: center;"><a href="https://github.com/THUDM/SwissArmyTransformer">SAT</a> FP16: 18GB <br><b>diffusers FP16: starting from 4GB*</b><br><b>diffusers INT8(torchao): starting from 3.6GB*</b></td>
113
+ <td style="text-align: center;"><a href="https://github.com/THUDM/SwissArmyTransformer">SAT</a> BF16: 26GB <br><b>diffusers BF16: starting from 5GB*</b><br><b>diffusers INT8(torchao): starting from 4.4GB*</b></td>
114
+ </tr>
115
+ <tr>
116
+ <td style="text-align: center;">Multi-GPU Inference VRAM Consumption</td>
117
+ <td style="text-align: center;"><b>FP16: 10GB* using diffusers</b></td>
118
+ <td style="text-align: center;"><b>BF16: 15GB* using diffusers</b></td>
119
+ </tr>
120
+ <tr>
121
+ <td style="text-align: center;">Inference Speed<br>(Step = 50, FP/BF16)</td>
122
+ <td style="text-align: center;">Single A100: ~90 seconds<br>Single H100: ~45 seconds</td>
123
+ <td style="text-align: center;">Single A100: ~180 seconds<br>Single H100: ~90 seconds</td>
124
+ </tr>
125
+ <tr>
126
+ <td style="text-align: center;">Fine-tuning Precision</td>
127
+ <td style="text-align: center;"><b>FP16</b></td>
128
+ <td style="text-align: center;"><b>BF16</b></td>
129
+ </tr>
130
+ <tr>
131
+ <td style="text-align: center;">Fine-tuning VRAM Consumption (per GPU)</td>
132
+ <td style="text-align: center;">47 GB (bs=1, LORA)<br> 61 GB (bs=2, LORA)<br> 62GB (bs=1, SFT)</td>
133
+ <td style="text-align: center;">63 GB (bs=1, LORA)<br> 80 GB (bs=2, LORA)<br> 75GB (bs=1, SFT)</td>
134
+ </tr>
135
+ <tr>
136
+ <td style="text-align: center;">Prompt Language</td>
137
+ <td colspan="2" style="text-align: center;">English*</td>
138
+ </tr>
139
+ <tr>
140
+ <td style="text-align: center;">Prompt Length Limit</td>
141
+ <td colspan="2" style="text-align: center;">226 Tokens</td>
142
+ </tr>
143
+ <tr>
144
+ <td style="text-align: center;">Video Length</td>
145
+ <td colspan="2" style="text-align: center;">6 Seconds</td>
146
+ </tr>
147
+ <tr>
148
+ <td style="text-align: center;">Frame Rate</td>
149
+ <td colspan="2" style="text-align: center;">8 Frames per Second</td>
150
+ </tr>
151
+ <tr>
152
+ <td style="text-align: center;">Video Resolution</td>
153
+ <td colspan="2" style="text-align: center;">720 x 480, no support for other resolutions (including fine-tuning)</td>
154
+ </tr>
155
+ <tr>
156
+ <td style="text-align: center;">Positional Encoding</td>
157
+ <td style="text-align: center;">3d_sincos_pos_embed</td>
158
+ <td style="text-align: center;">3d_rope_pos_embed</td>
159
+ </tr>
160
+ </table>
161
+
162
+ **Data Explanation**
163
+
164
+ + When testing using the `diffusers` library, all optimizations provided by the `diffusers` library were enabled. This
165
+ solution has not been tested for actual VRAM/memory usage on devices other than **NVIDIA A100 / H100**. Generally,
166
+ this solution can be adapted to all devices with **NVIDIA Ampere architecture** and above. If the optimizations are
167
+ disabled, VRAM usage will increase significantly, with peak VRAM usage being about 3 times higher than the table
168
+ shows. However, speed will increase by 3-4 times. You can selectively disable some optimizations, including:
169
+
170
+ ```
171
+ pipe.enable_model_cpu_offload()
172
+ pipe.enable_sequential_cpu_offload()
173
+ pipe.vae.enable_slicing()
174
+ pipe.vae.enable_tiling()
175
+ ```
176
+
177
+ + When performing multi-GPU inference, the `enable_model_cpu_offload()` optimization needs to be disabled.
178
+ + Using INT8 models will reduce inference speed. This is to ensure that GPUs with lower VRAM can perform inference
179
+ normally while maintaining minimal video quality loss, though inference speed will decrease significantly.
180
+ + The 2B model is trained with `FP16` precision, and the 5B model is trained with `BF16` precision. We recommend using
181
+ the precision the model was trained with for inference.
182
+ + [PytorchAO](https://github.com/pytorch/ao) and [Optimum-quanto](https://github.com/huggingface/optimum-quanto/) can be
183
+ used to quantize the text encoder, Transformer, and VAE modules to reduce CogVideoX's memory requirements. This makes
184
+ it possible to run the model on a free T4 Colab or GPUs with smaller VRAM! It is also worth noting that TorchAO
185
+ quantization is fully compatible with `torch.compile`, which can significantly improve inference speed. `FP8`
186
+ precision must be used on devices with `NVIDIA H100` or above, which requires installing
187
+ the `torch`, `torchao`, `diffusers`, and `accelerate` Python packages from source. `CUDA 12.4` is recommended.
188
+ + The inference speed test also used the above VRAM optimization scheme. Without VRAM optimization, inference speed
189
+ increases by about 10%. Only the `diffusers` version of the model supports quantization.
190
+ + The model only supports English input; other languages can be translated into English during refinement by a large
191
+ model.
192
+
193
+ **Note**
194
+
195
+ + Using [SAT](https://github.com/THUDM/SwissArmyTransformer) for inference and fine-tuning of SAT version
196
+ models. Feel free to visit our GitHub for more information.
197
+
198
+ ## Quick Start 🤗
199
+
200
+ This model supports deployment using the huggingface diffusers library. You can deploy it by following these steps.
201
+
202
+ **We recommend that you visit our [GitHub](https://github.com/THUDM/CogVideo) and check out the relevant prompt
203
+ optimizations and conversions to get a better experience.**
204
+
205
+ 1. Install the required dependencies
206
+
207
+ ```shell
208
+ # diffusers>=0.30.1
209
+ # transformers>=0.44.0
210
+ # accelerate>=0.33.0 (suggest install from source)
211
+ # imageio-ffmpeg>=0.5.1
212
+ pip install --upgrade transformers accelerate diffusers imageio-ffmpeg
213
+ ```
214
+
215
+ 2. Run the code
216
+
217
+ ```python
218
+ import torch
219
+ from diffusers import CogVideoXPipeline
220
+ from diffusers.utils import export_to_video
221
+
222
+ prompt = "A panda, dressed in a small, red jacket and a tiny hat, sits on a wooden stool in a serene bamboo forest. The panda's fluffy paws strum a miniature acoustic guitar, producing soft, melodic tunes. Nearby, a few other pandas gather, watching curiously and some clapping in rhythm. Sunlight filters through the tall bamboo, casting a gentle glow on the scene. The panda's face is expressive, showing concentration and joy as it plays. The background includes a small, flowing stream and vibrant green foliage, enhancing the peaceful and magical atmosphere of this unique musical performance."
223
+
224
+ pipe = CogVideoXPipeline.from_pretrained(
225
+ "THUDM/CogVideoX-2b",
226
+ torch_dtype=torch.float16
227
+ )
228
+
229
+ pipe.enable_model_cpu_offload()
230
+ pipe.enable_sequential_cpu_offload()
231
+ pipe.vae.enable_slicing()
232
+ pipe.vae.enable_tiling()
233
+ video = pipe(
234
+ prompt=prompt,
235
+ num_videos_per_prompt=1,
236
+ num_inference_steps=50,
237
+ num_frames=49,
238
+ guidance_scale=6,
239
+ generator=torch.Generator(device="cuda").manual_seed(42),
240
+ ).frames[0]
241
+
242
+ export_to_video(video, "output.mp4", fps=8)
243
+ ```
244
+
245
+ ## Quantized Inference
246
+
247
+ [PytorchAO](https://github.com/pytorch/ao) and [Optimum-quanto](https://github.com/huggingface/optimum-quanto/) can be
248
+ used to quantize the Text Encoder, Transformer and VAE modules to lower the memory requirement of CogVideoX. This makes
249
+ it possible to run the model on free-tier T4 Colab or smaller VRAM GPUs as well! It is also worth noting that TorchAO
250
+ quantization is fully compatible with `torch.compile`, which allows for much faster inference speed.
251
+
252
+ ```diff
253
+ # To get started, PytorchAO needs to be installed from the GitHub source and PyTorch Nightly.
254
+ # Source and nightly installation is only required until next release.
255
+
256
+ import torch
257
+ from diffusers import AutoencoderKLCogVideoX, CogVideoXTransformer3DModel, CogVideoXPipeline
258
+ from diffusers.utils import export_to_video
259
+ + from transformers import T5EncoderModel
260
+ + from torchao.quantization import quantize_, int8_weight_only, int8_dynamic_activation_int8_weight
261
+
262
+ + quantization = int8_weight_only
263
+
264
+ + text_encoder = T5EncoderModel.from_pretrained("THUDM/CogVideoX-5b", subfolder="text_encoder", torch_dtype=torch.bfloat16)
265
+ + quantize_(text_encoder, quantization())
266
+
267
+ + transformer = CogVideoXTransformer3DModel.from_pretrained("THUDM/CogVideoX-5b", subfolder="transformer", torch_dtype=torch.bfloat16)
268
+ + quantize_(transformer, quantization())
269
+
270
+ + vae = AutoencoderKLCogVideoX.from_pretrained("THUDM/CogVideoX-2b", subfolder="vae", torch_dtype=torch.bfloat16)
271
+ + quantize_(vae, quantization())
272
+
273
+ # Create pipeline and run inference
274
+ pipe = CogVideoXPipeline.from_pretrained(
275
+ "THUDM/CogVideoX-2b",
276
+ + text_encoder=text_encoder,
277
+ + transformer=transformer,
278
+ + vae=vae,
279
+ torch_dtype=torch.bfloat16,
280
+ )
281
+ pipe.enable_model_cpu_offload()
282
+ pipe.vae.enable_tiling()
283
+
284
+ prompt = "A panda, dressed in a small, red jacket and a tiny hat, sits on a wooden stool in a serene bamboo forest. The panda's fluffy paws strum a miniature acoustic guitar, producing soft, melodic tunes. Nearby, a few other pandas gather, watching curiously and some clapping in rhythm. Sunlight filters through the tall bamboo, casting a gentle glow on the scene. The panda's face is expressive, showing concentration and joy as it plays. The background includes a small, flowing stream and vibrant green foliage, enhancing the peaceful and magical atmosphere of this unique musical performance."
285
+
286
+ video = pipe(
287
+ prompt=prompt,
288
+ num_videos_per_prompt=1,
289
+ num_inference_steps=50,
290
+ num_frames=49,
291
+ guidance_scale=6,
292
+ generator=torch.Generator(device="cuda").manual_seed(42),
293
+ ).frames[0]
294
+
295
+ export_to_video(video, "output.mp4", fps=8)
296
+ ```
297
+
298
+ Additionally, the models can be serialized and stored in a quantized datatype to save disk space when using PytorchAO.
299
+ Find examples and benchmarks at these links:
300
+
301
+ - [torchao](https://gist.github.com/a-r-r-o-w/4d9732d17412888c885480c6521a9897)
302
+ - [quanto](https://gist.github.com/a-r-r-o-w/31be62828b00a9292821b85c1017effa)
303
+
304
+ ## Explore the Model
305
+
306
+ Welcome to our [github](https://github.com/THUDM/CogVideo), where you will find:
307
+
308
+ 1. More detailed technical details and code explanation.
309
+ 2. Optimization and conversion of prompt words.
310
+ 3. Reasoning and fine-tuning of SAT version models, and even pre-release.
311
+ 4. Project update log dynamics, more interactive opportunities.
312
+ 5. CogVideoX toolchain to help you better use the model.
313
+ 6. INT8 model inference code support.
314
+
315
+ ## Model License
316
+
317
+ The CogVideoX-2B model (including its corresponding Transformers module and VAE module) is released under
318
+ the [Apache 2.0 License](LICENSE).
319
+
320
+ The CogVideoX-5B model (Transformers module) is released under
321
+ the [CogVideoX LICENSE](https://huggingface.co/THUDM/CogVideoX-5b/blob/main/LICENSE).
322
+
323
+ ## Citation
324
+
325
+ ```
326
+ @article{yang2024cogvideox,
327
+ title={CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer},
328
+ author={Yang, Zhuoyi and Teng, Jiayan and Zheng, Wendi and Ding, Ming and Huang, Shiyu and Xu, Jiazheng and Yang, Yuanming and Hong, Wenyi and Zhang, Xiaohan and Feng, Guanyu and others},
329
+ journal={arXiv preprint arXiv:2408.06072},
330
+ year={2024}
331
+ }
332
+ ```
README_zh.md ADDED
@@ -0,0 +1,305 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # CogVideoX-2B
2
+
3
+ <p style="text-align: center;">
4
+ <div align="center">
5
+ <img src=https://github.com/THUDM/CogVideo/raw/main/resources/logo.svg width="50%"/>
6
+ </div>
7
+ <p align="center">
8
+ <a href="https://huggingface.co/THUDM/CogVideoX-2b/blob/main/README.md">📄 Read in English</a> |
9
+ <a href="https://huggingface.co/spaces/THUDM/CogVideoX-2B-Space">🤗 Huggingface Space</a> |
10
+ <a href="https://github.com/THUDM/CogVideo">🌐 Github </a> |
11
+ <a href="https://arxiv.org/pdf/2408.06072">📜 arxiv </a>
12
+ </p>
13
+ <p align="center">
14
+ 📍 前往<a href="https://chatglm.cn/video?fr=osm_cogvideox"> 清影</a> 和 <a href="https://open.bigmodel.cn/?utm_campaign=open&_channel_track_key=OWTVNma9"> API平台</a> 体验商业版视频生成模型
15
+ </p>
16
+
17
+ ## 作品案例
18
+
19
+ <!DOCTYPE html>
20
+ <html lang="en">
21
+ <head>
22
+ <meta charset="UTF-8">
23
+ <meta name="viewport" content="width=device-width, initial-scale=1.0">
24
+ <title>Video Gallery with Captions</title>
25
+ <style>
26
+ .video-container {
27
+ display: flex;
28
+ flex-wrap: wrap;
29
+ justify-content: space-around;
30
+ }
31
+ .video-item {
32
+ width: 45%;
33
+ margin-bottom: 20px;
34
+ transition: transform 0.3s;
35
+ }
36
+ .video-item:hover {
37
+ transform: scale(1.1);
38
+ }
39
+ .caption {
40
+ text-align: center;
41
+ margin-top: 10px;
42
+ font-size: 11px;
43
+ }
44
+ </style>
45
+ </head>
46
+ <body>
47
+ <div class="video-container">
48
+ <div class="video-item">
49
+ <video width="100%" controls>
50
+ <source src="https://github.com/THUDM/CogVideo/raw/main/resources/videos/1.mp4" type="video/mp4">
51
+ </video>
52
+ <div class="caption">A detailed wooden toy ship with intricately carved masts and sails is seen gliding smoothly over a plush, blue carpet that mimics the waves of the sea. The ship's hull is painted a rich brown, with tiny windows. The carpet, soft and textured, provides a perfect backdrop, resembling an oceanic expanse. Surrounding the ship are various other toys and children's items, hinting at a playful environment. The scene captures the innocence and imagination of childhood, with the toy ship's journey symbolizing endless adventures in a whimsical, indoor setting.</div>
53
+ </div>
54
+ <div class="video-item">
55
+ <video width="100%" controls>
56
+ <source src="https://github.com/THUDM/CogVideo/raw/main/resources/videos/2.mp4" type="video/mp4">
57
+ </video>
58
+ <div class="caption">The camera follows behind a white vintage SUV with a black roof rack as it speeds up a steep dirt road surrounded by pine trees on a steep mountain slope, dust kicks up from it’s tires, the sunlight shines on the SUV as it speeds along the dirt road, casting a warm glow over the scene. The dirt road curves gently into the distance, with no other cars or vehicles in sight. The trees on either side of the road are redwoods, with patches of greenery scattered throughout. The car is seen from the rear following the curve with ease, making it seem as if it is on a rugged drive through the rugged terrain. The dirt road itself is surrounded by steep hills and mountains, with a clear blue sky above with wispy clouds.</div>
59
+ </div>
60
+ <div class="video-item">
61
+ <video width="100%" controls>
62
+ <source src="https://github.com/THUDM/CogVideo/raw/main/resources/videos/3.mp4" type="video/mp4">
63
+ </video>
64
+ <div class="caption">A street artist, clad in a worn-out denim jacket and a colorful bandana, stands before a vast concrete wall in the heart, holding a can of spray paint, spray-painting a colorful bird on a mottled wall.</div>
65
+ </div>
66
+ <div class="video-item">
67
+ <video width="100%" controls>
68
+ <source src="https://github.com/THUDM/CogVideo/raw/main/resources/videos/4.mp4" type="video/mp4">
69
+ </video>
70
+ <div class="caption"> In the haunting backdrop of a war-torn city, where ruins and crumbled walls tell a story of devastation, a poignant close-up frames a young girl. Her face is smudged with ash, a silent testament to the chaos around her. Her eyes glistening with a mix of sorrow and resilience, capturing the raw emotion of a world that has lost its innocence to the ravages of conflict.</div>
71
+ </div>
72
+ </div>
73
+ </body>
74
+ </html>
75
+
76
+ ## 模型介绍
77
+
78
+ CogVideoX是 [清影](https://chatglm.cn/video?fr=osm_cogvideo) 同源的开源版本视频生成模型。下表展示目前我们提供的视频生成模型列表,以及相关基础信息。
79
+
80
+ <table style="border-collapse: collapse; width: 100%;">
81
+ <tr>
82
+ <th style="text-align: center;">模型名</th>
83
+ <th style="text-align: center;">CogVideoX-2B (本仓库)</th>
84
+ <th style="text-align: center;">CogVideoX-5B </th>
85
+ </tr>
86
+ <tr>
87
+ <td style="text-align: center;">模型介绍</td>
88
+ <td style="text-align: center;">入门级模型,兼顾兼容性��运行,二次开发成本低。</td>
89
+ <td style="text-align: center;">视频生成质量更高,视觉效果更好的更大尺寸模型。</td>
90
+ </tr>
91
+ <tr>
92
+ <td style="text-align: center;">推理精度</td>
93
+ <td style="text-align: center;"><b>FP16*(推荐)</b>, BF16, FP32,FP8*,INT8,不支持INT4</td>
94
+ <td style="text-align: center;"><b>BF16(推荐)</b>, FP16, FP32,FP8*,INT8,不支持INT4</td>
95
+ </tr>
96
+ <tr>
97
+ <td style="text-align: center;">单GPU显存消耗<br></td>
98
+ <td style="text-align: center;"><a href="https://github.com/THUDM/SwissArmyTransformer">SAT</a> FP16: 18GB <br><b>diffusers FP16: 4GB起* </b><br><b>diffusers INT8(torchao): 3.6G起*</b></td>
99
+ <td style="text-align: center;"><a href="https://github.com/THUDM/SwissArmyTransformer">SAT</a> BF16: 26GB <br><b>diffusers BF16 : 5GB起* </b><br><b>diffusers INT8(torchao): 4.4G起* </b></td>
100
+ </tr>
101
+ <tr>
102
+ <td style="text-align: center;">多GPU推理显存消耗</td>
103
+ <td style="text-align: center;"><b>FP16: 10GB* using diffusers</b><br></td>
104
+ <td style="text-align: center;"><b>BF16: 15GB* using diffusers</b><br></td>
105
+ </tr>
106
+ <tr>
107
+ <td style="text-align: center;">推理速度<br>(Step = 50, FP/BF16)</td>
108
+ <td style="text-align: center;">单卡A100: ~90秒<br>单卡H100: ~45秒</td>
109
+ <td style="text-align: center;">单卡A100: ~180秒<br>单卡H100: ~90秒</td>
110
+ </tr>
111
+ <tr>
112
+ <td style="text-align: center;">微调精度</td>
113
+ <td style="text-align: center;"><b>FP16</b></td>
114
+ <td style="text-align: center;"><b>BF16</b></td>
115
+ </tr>
116
+ <tr>
117
+ <td style="text-align: center;">微调显存消耗(每卡)</td>
118
+ <td style="text-align: center;">47 GB (bs=1, LORA)<br> 61 GB (bs=2, LORA)<br> 62GB (bs=1, SFT)</td>
119
+ <td style="text-align: center;">63 GB (bs=1, LORA)<br> 80 GB (bs=2, LORA)<br> 75GB (bs=1, SFT)<br></td>
120
+ </tr>
121
+ <tr>
122
+ <td style="text-align: center;">提示词语言</td>
123
+ <td colspan="2" style="text-align: center;">English*</td>
124
+ </tr>
125
+ <tr>
126
+ <td style="text-align: center;">提示词长度上限</td>
127
+ <td colspan="2" style="text-align: center;">226 Tokens</td>
128
+ </tr>
129
+ <tr>
130
+ <td style="text-align: center;">视频长度</td>
131
+ <td colspan="2" style="text-align: center;">6 秒</td>
132
+ </tr>
133
+ <tr>
134
+ <td style="text-align: center;">帧率</td>
135
+ <td colspan="2" style="text-align: center;">8 帧 / 秒 </td>
136
+ </tr>
137
+ <tr>
138
+ <td style="text-align: center;">视频分辨率</td>
139
+ <td colspan="2" style="text-align: center;">720 * 480,不支持其他分辨率(含微调)</td>
140
+ </tr>
141
+ <tr>
142
+ <td style="text-align: center;">位置编码</td>
143
+ <td style="text-align: center;">3d_sincos_pos_embed</td>
144
+ <td style="text-align: center;">3d_rope_pos_embed<br></td>
145
+ </tr>
146
+ </table>
147
+
148
+ **数据解释**
149
+
150
+ + 使用 diffusers 库进行测试时,启用了全部`diffusers`库自带的优化,该方案未测试在非**NVIDIA A100 / H100** 外的设备上的实际显存 / 内存占用。通常,该方案可以适配于所有 **NVIDIA 安培架构**
151
+ 以上的设备。若关闭优化,显存占用会成倍增加,峰值显存约为表格的3倍。但速度提升3-4倍左右。你可以选择性的关闭部分优化,这些优化包括:
152
+ ```
153
+ pipe.enable_model_cpu_offload()
154
+ pipe.enable_sequential_cpu_offload()
155
+ pipe.vae.enable_slicing()
156
+ pipe.vae.enable_tiling()
157
+ ```
158
+
159
+ + 多GPU推理时,需要关闭 `enable_model_cpu_offload()` 优化。
160
+ + 使用 INT8 模型会导致推理速度降低,此举是为了满足显存较低的显卡能正常推理并保持较少的视频质量损失,推理速度大幅降低。
161
+ + 2B 模型采用 `FP16` 精度训练, 5B模型采用 `BF16` 精度训练。我们推荐使用模型训练的精度进行推理。
162
+ + [PytorchAO](https://github.com/pytorch/ao) 和 [Optimum-quanto](https://github.com/huggingface/optimum-quanto/)
163
+ 可以用于量化文本编码器、Transformer 和 VAE 模块,以降低 CogVideoX 的内存需求。这使得在免费的 T4 Colab 或更小显存的 GPU
164
+ 上运行模型成为可能!同样值得注意的是,TorchAO 量化完全兼容 `torch.compile`,这可以显著提高推理速度。在 `NVIDIA H100`
165
+ 及以上设备上必须使用 `FP8` 精度,这需要源码安装 `torch`、`torchao`、`diffusers` 和 `accelerate` Python
166
+ 包。建议使用 `CUDA 12.4`。
167
+ + 推理速度测试同样采用了上述显存优化方案,不采用显存优化的情况下,推理速度提升约10%。 只有`diffusers`版本模型支持量化。
168
+ + 模型仅支持英语输入,其他语言可以通过大模型润色时翻译为英语。
169
+
170
+ **提醒**
171
+
172
+ + 使用 [SAT](https://github.com/THUDM/SwissArmyTransformer) 推理和微调SAT版本模型。欢迎前往我们的github查看。
173
+
174
+ ## 快速上手 🤗
175
+
176
+ 本模型已经支持使用 huggingface 的 diffusers 库进行部署,你可以按照以下步骤进行部署。
177
+
178
+ **我们推荐您进入我们的 [github](https://github.com/THUDM/CogVideo) 并查看相关的提示词优化和转换,以获得更���的体验。**
179
+
180
+ 1. 安装对应的依赖
181
+
182
+ ```shell
183
+ # diffusers>=0.30.1
184
+ # transformers>=0.44.0
185
+ # accelerate>=0.33.0 (suggest install from source)
186
+ # imageio-ffmpeg>=0.5.1
187
+ pip install --upgrade transformers accelerate diffusers imageio-ffmpeg
188
+ ```
189
+
190
+ 2. 运行代码 (BF16 / FP16)
191
+
192
+ ```python
193
+ import torch
194
+ from diffusers import CogVideoXPipeline
195
+ from diffusers.utils import export_to_video
196
+
197
+ prompt = "A panda, dressed in a small, red jacket and a tiny hat, sits on a wooden stool in a serene bamboo forest. The panda's fluffy paws strum a miniature acoustic guitar, producing soft, melodic tunes. Nearby, a few other pandas gather, watching curiously and some clapping in rhythm. Sunlight filters through the tall bamboo, casting a gentle glow on the scene. The panda's face is expressive, showing concentration and joy as it plays. The background includes a small, flowing stream and vibrant green foliage, enhancing the peaceful and magical atmosphere of this unique musical performance."
198
+
199
+ pipe = CogVideoXPipeline.from_pretrained(
200
+ "THUDM/CogVideoX-2b",
201
+ torch_dtype=torch.float16
202
+ )
203
+
204
+ pipe.enable_model_cpu_offload()
205
+ pipe.enable_sequential_cpu_offload()
206
+ pipe.vae.enable_slicing()
207
+ pipe.vae.enable_tiling()
208
+ video = pipe(
209
+ prompt=prompt,
210
+ num_videos_per_prompt=1,
211
+ num_inference_steps=50,
212
+ num_frames=49,
213
+ guidance_scale=6,
214
+ generator=torch.Generator(device="cuda").manual_seed(42),
215
+ ).frames[0]
216
+
217
+ export_to_video(video, "output.mp4", fps=8)
218
+ ```
219
+
220
+ ## Quantized Inference
221
+
222
+ [PytorchAO](https://github.com/pytorch/ao) 和 [Optimum-quanto](https://github.com/huggingface/optimum-quanto/)
223
+ 可以用于对文本编码器、Transformer 和 VAE 模块进行量化,从而降低 CogVideoX 的内存需求。这使得在免费的 T4 Colab 或较小 VRAM 的
224
+ GPU 上运行该模型成为可能!值得注意的是,TorchAO 量化与 `torch.compile` 完全兼容,这可以显著加快推理速度。
225
+
226
+ ```diff
227
+ # To get started, PytorchAO needs to be installed from the GitHub source and PyTorch Nightly.
228
+ # Source and nightly installation is only required until next release.
229
+
230
+ import torch
231
+ from diffusers import AutoencoderKLCogVideoX, CogVideoXTransformer3DModel, CogVideoXPipeline
232
+ from diffusers.utils import export_to_video
233
+ + from transformers import T5EncoderModel
234
+ + from torchao.quantization import quantize_, int8_weight_only, int8_dynamic_activation_int8_weight
235
+
236
+ + quantization = int8_weight_only
237
+
238
+ + text_encoder = T5EncoderModel.from_pretrained("THUDM/CogVideoX-2b", subfolder="text_encoder", torch_dtype=torch.bfloat16)
239
+ + quantize_(text_encoder, quantization())
240
+
241
+ + transformer = CogVideoXTransformer3DModel.from_pretrained("THUDM/CogVideoX-5b", subfolder="transformer", torch_dtype=torch.bfloat16)
242
+ + quantize_(transformer, quantization())
243
+
244
+ + vae = AutoencoderKLCogVideoX.from_pretrained("THUDM/CogVideoX-2b", subfolder="vae", torch_dtype=torch.bfloat16)
245
+ + quantize_(vae, quantization())
246
+
247
+ # Create pipeline and run inference
248
+ pipe = CogVideoXPipeline.from_pretrained(
249
+ "THUDM/CogVideoX-2b",
250
+ + text_encoder=text_encoder,
251
+ + transformer=transformer,
252
+ + vae=vae,
253
+ torch_dtype=torch.bfloat16,
254
+ )
255
+ pipe.enable_model_cpu_offload()
256
+ pipe.vae.enable_tiling()
257
+
258
+ prompt = "A panda, dressed in a small, red jacket and a tiny hat, sits on a wooden stool in a serene bamboo forest. The panda's fluffy paws strum a miniature acoustic guitar, producing soft, melodic tunes. Nearby, a few other pandas gather, watching curiously and some clapping in rhythm. Sunlight filters through the tall bamboo, casting a gentle glow on the scene. The panda's face is expressive, showing concentration and joy as it plays. The background includes a small, flowing stream and vibrant green foliage, enhancing the peaceful and magical atmosphere of this unique musical performance."
259
+
260
+ video = pipe(
261
+ prompt=prompt,
262
+ num_videos_per_prompt=1,
263
+ num_inference_steps=50,
264
+ num_frames=49,
265
+ guidance_scale=6,
266
+ generator=torch.Generator(device="cuda").manual_seed(42),
267
+ ).frames[0]
268
+
269
+ export_to_video(video, "output.mp4", fps=8)
270
+ ```
271
+
272
+ 此外,这些模型可以通过使用PytorchAO以量化数据类型序列化并存储,从而节省磁盘空间。你可以在以下链接中找到示例和基准测试。
273
+
274
+ - [torchao](https://gist.github.com/a-r-r-o-w/4d9732d17412888c885480c6521a9897)
275
+ - [quanto](https://gist.github.com/a-r-r-o-w/31be62828b00a9292821b85c1017effa)
276
+
277
+ ## 深入研究
278
+
279
+ 欢迎进入我们的 [github](https://github.com/THUDM/CogVideo),你将获得:
280
+
281
+ 1. 更加详细的技术细节介绍和代码解释。
282
+ 2. 提示词的优化和转换。
283
+ 3. SAT版本模型进行推理和微调,甚至预发布。
284
+ 4. 项目更新日志动态,更多互动机会。
285
+ 5. CogVideoX 工具链,帮助您更好的使用模型。
286
+ 6. INT8 模型推理代码。
287
+
288
+ ## 模型协议
289
+
290
+ CogVideoX-2B 模型 (包括其对应的Transformers模块,VAE模块) 根据 [Apache 2.0 License](LICENSE) 许可证发布。
291
+
292
+ CogVideoX-5B 模型 (Transformers 模块)
293
+ 根据 [CogVideoX LICENSE](https://huggingface.co/THUDM/CogVideoX-5b/blob/main/LICENSE)
294
+ 许可证发布。
295
+
296
+ ## 引用
297
+
298
+ ```
299
+ @article{yang2024cogvideox,
300
+ title={CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer},
301
+ author={Yang, Zhuoyi and Teng, Jiayan and Zheng, Wendi and Ding, Ming and Huang, Shiyu and Xu, Jiazheng and Yang, Yuanming and Hong, Wenyi and Zhang, Xiaohan and Feng, Guanyu and others},
302
+ journal={arXiv preprint arXiv:2408.06072},
303
+ year={2024}
304
+ }
305
+ ```
model_index.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_class_name": "CogVideoXPipeline",
3
+ "_diffusers_version": "0.30.0.dev0",
4
+ "scheduler": [
5
+ "diffusers",
6
+ "CogVideoXDDIMScheduler"
7
+ ],
8
+ "text_encoder": [
9
+ "transformers",
10
+ "T5EncoderModel"
11
+ ],
12
+ "tokenizer": [
13
+ "transformers",
14
+ "T5Tokenizer"
15
+ ],
16
+ "transformer": [
17
+ "diffusers",
18
+ "CogVideoXTransformer3DModel"
19
+ ],
20
+ "vae": [
21
+ "diffusers",
22
+ "AutoencoderKLCogVideoX"
23
+ ]
24
+ }
scheduler/scheduler_config.json ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_class_name": "CogVideoXDDIMScheduler",
3
+ "_diffusers_version": "0.30.0.dev0",
4
+ "beta_end": 0.012,
5
+ "beta_schedule": "scaled_linear",
6
+ "beta_start": 0.00085,
7
+ "clip_sample": false,
8
+ "clip_sample_range": 1.0,
9
+ "num_train_timesteps": 1000,
10
+ "prediction_type": "v_prediction",
11
+ "rescale_betas_zero_snr": true,
12
+ "sample_max_value": 1.0,
13
+ "set_alpha_to_one": true,
14
+ "snr_shift_scale": 3.0,
15
+ "steps_offset": 0,
16
+ "timestep_spacing": "trailing",
17
+ "trained_betas": null
18
+ }
text_encoder/config.json ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "google/t5-v1_1-xxl_1",
3
+ "architectures": [
4
+ "T5EncoderModel"
5
+ ],
6
+ "classifier_dropout": 0.0,
7
+ "d_ff": 10240,
8
+ "d_kv": 64,
9
+ "d_model": 4096,
10
+ "decoder_start_token_id": 0,
11
+ "dense_act_fn": "gelu_new",
12
+ "dropout_rate": 0.1,
13
+ "eos_token_id": 1,
14
+ "feed_forward_proj": "gated-gelu",
15
+ "initializer_factor": 1.0,
16
+ "is_encoder_decoder": true,
17
+ "is_gated_act": true,
18
+ "layer_norm_epsilon": 1e-06,
19
+ "model_type": "t5",
20
+ "num_decoder_layers": 24,
21
+ "num_heads": 64,
22
+ "num_layers": 24,
23
+ "output_past": true,
24
+ "pad_token_id": 0,
25
+ "relative_attention_max_distance": 128,
26
+ "relative_attention_num_buckets": 32,
27
+ "tie_word_embeddings": false,
28
+ "torch_dtype": "float16",
29
+ "transformers_version": "4.43.2",
30
+ "use_cache": true,
31
+ "vocab_size": 32128
32
+ }
text_encoder/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4f2751ceeb2a96edd693e539dc5d6bba0b8d3814f49a9b3798403a0cec4b2e3d
3
+ size 4994582104
text_encoder/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f63154532130422309532ff56f11945fbea8266c958e3133e8e5aef85c6293c7
3
+ size 4530066248
text_encoder/model.safetensors.index.json ADDED
@@ -0,0 +1,226 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 9524621312
4
+ },
5
+ "weight_map": {
6
+ "encoder.block.0.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
7
+ "encoder.block.0.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
8
+ "encoder.block.0.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
9
+ "encoder.block.0.layer.0.SelfAttention.relative_attention_bias.weight": "model-00001-of-00002.safetensors",
10
+ "encoder.block.0.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
11
+ "encoder.block.0.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
12
+ "encoder.block.0.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
13
+ "encoder.block.0.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
14
+ "encoder.block.0.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
15
+ "encoder.block.0.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
16
+ "encoder.block.1.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
17
+ "encoder.block.1.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
18
+ "encoder.block.1.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
19
+ "encoder.block.1.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
20
+ "encoder.block.1.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
21
+ "encoder.block.1.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
22
+ "encoder.block.1.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
23
+ "encoder.block.1.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
24
+ "encoder.block.1.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
25
+ "encoder.block.10.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
26
+ "encoder.block.10.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
27
+ "encoder.block.10.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
28
+ "encoder.block.10.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
29
+ "encoder.block.10.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
30
+ "encoder.block.10.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
31
+ "encoder.block.10.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
32
+ "encoder.block.10.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
33
+ "encoder.block.10.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
34
+ "encoder.block.11.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
35
+ "encoder.block.11.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
36
+ "encoder.block.11.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
37
+ "encoder.block.11.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
38
+ "encoder.block.11.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
39
+ "encoder.block.11.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
40
+ "encoder.block.11.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
41
+ "encoder.block.11.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
42
+ "encoder.block.11.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
43
+ "encoder.block.12.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
44
+ "encoder.block.12.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
45
+ "encoder.block.12.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
46
+ "encoder.block.12.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
47
+ "encoder.block.12.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
48
+ "encoder.block.12.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
49
+ "encoder.block.12.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
50
+ "encoder.block.12.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
51
+ "encoder.block.12.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
52
+ "encoder.block.13.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
53
+ "encoder.block.13.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
54
+ "encoder.block.13.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
55
+ "encoder.block.13.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
56
+ "encoder.block.13.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
57
+ "encoder.block.13.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
58
+ "encoder.block.13.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
59
+ "encoder.block.13.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
60
+ "encoder.block.13.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
61
+ "encoder.block.14.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
62
+ "encoder.block.14.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
63
+ "encoder.block.14.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
64
+ "encoder.block.14.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
65
+ "encoder.block.14.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
66
+ "encoder.block.14.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
67
+ "encoder.block.14.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
68
+ "encoder.block.14.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
69
+ "encoder.block.14.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
70
+ "encoder.block.15.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
71
+ "encoder.block.15.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
72
+ "encoder.block.15.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
73
+ "encoder.block.15.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
74
+ "encoder.block.15.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
75
+ "encoder.block.15.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
76
+ "encoder.block.15.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
77
+ "encoder.block.15.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
78
+ "encoder.block.15.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
79
+ "encoder.block.16.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
80
+ "encoder.block.16.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
81
+ "encoder.block.16.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
82
+ "encoder.block.16.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
83
+ "encoder.block.16.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
84
+ "encoder.block.16.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
85
+ "encoder.block.16.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
86
+ "encoder.block.16.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
87
+ "encoder.block.16.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
88
+ "encoder.block.17.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
89
+ "encoder.block.17.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
90
+ "encoder.block.17.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
91
+ "encoder.block.17.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
92
+ "encoder.block.17.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
93
+ "encoder.block.17.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
94
+ "encoder.block.17.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
95
+ "encoder.block.17.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
96
+ "encoder.block.17.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
97
+ "encoder.block.18.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
98
+ "encoder.block.18.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
99
+ "encoder.block.18.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
100
+ "encoder.block.18.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
101
+ "encoder.block.18.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
102
+ "encoder.block.18.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
103
+ "encoder.block.18.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
104
+ "encoder.block.18.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
105
+ "encoder.block.18.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
106
+ "encoder.block.19.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
107
+ "encoder.block.19.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
108
+ "encoder.block.19.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
109
+ "encoder.block.19.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
110
+ "encoder.block.19.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
111
+ "encoder.block.19.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
112
+ "encoder.block.19.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
113
+ "encoder.block.19.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
114
+ "encoder.block.19.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
115
+ "encoder.block.2.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
116
+ "encoder.block.2.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
117
+ "encoder.block.2.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
118
+ "encoder.block.2.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
119
+ "encoder.block.2.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
120
+ "encoder.block.2.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
121
+ "encoder.block.2.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
122
+ "encoder.block.2.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
123
+ "encoder.block.2.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
124
+ "encoder.block.20.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
125
+ "encoder.block.20.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
126
+ "encoder.block.20.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
127
+ "encoder.block.20.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
128
+ "encoder.block.20.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
129
+ "encoder.block.20.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
130
+ "encoder.block.20.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
131
+ "encoder.block.20.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
132
+ "encoder.block.20.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
133
+ "encoder.block.21.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
134
+ "encoder.block.21.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
135
+ "encoder.block.21.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
136
+ "encoder.block.21.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
137
+ "encoder.block.21.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
138
+ "encoder.block.21.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
139
+ "encoder.block.21.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
140
+ "encoder.block.21.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
141
+ "encoder.block.21.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
142
+ "encoder.block.22.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
143
+ "encoder.block.22.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
144
+ "encoder.block.22.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
145
+ "encoder.block.22.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
146
+ "encoder.block.22.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
147
+ "encoder.block.22.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
148
+ "encoder.block.22.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
149
+ "encoder.block.22.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
150
+ "encoder.block.22.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
151
+ "encoder.block.23.layer.0.SelfAttention.k.weight": "model-00002-of-00002.safetensors",
152
+ "encoder.block.23.layer.0.SelfAttention.o.weight": "model-00002-of-00002.safetensors",
153
+ "encoder.block.23.layer.0.SelfAttention.q.weight": "model-00002-of-00002.safetensors",
154
+ "encoder.block.23.layer.0.SelfAttention.v.weight": "model-00002-of-00002.safetensors",
155
+ "encoder.block.23.layer.0.layer_norm.weight": "model-00002-of-00002.safetensors",
156
+ "encoder.block.23.layer.1.DenseReluDense.wi_0.weight": "model-00002-of-00002.safetensors",
157
+ "encoder.block.23.layer.1.DenseReluDense.wi_1.weight": "model-00002-of-00002.safetensors",
158
+ "encoder.block.23.layer.1.DenseReluDense.wo.weight": "model-00002-of-00002.safetensors",
159
+ "encoder.block.23.layer.1.layer_norm.weight": "model-00002-of-00002.safetensors",
160
+ "encoder.block.3.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
161
+ "encoder.block.3.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
162
+ "encoder.block.3.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
163
+ "encoder.block.3.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
164
+ "encoder.block.3.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
165
+ "encoder.block.3.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
166
+ "encoder.block.3.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
167
+ "encoder.block.3.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
168
+ "encoder.block.3.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
169
+ "encoder.block.4.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
170
+ "encoder.block.4.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
171
+ "encoder.block.4.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
172
+ "encoder.block.4.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
173
+ "encoder.block.4.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
174
+ "encoder.block.4.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
175
+ "encoder.block.4.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
176
+ "encoder.block.4.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
177
+ "encoder.block.4.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
178
+ "encoder.block.5.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
179
+ "encoder.block.5.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
180
+ "encoder.block.5.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
181
+ "encoder.block.5.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
182
+ "encoder.block.5.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
183
+ "encoder.block.5.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
184
+ "encoder.block.5.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
185
+ "encoder.block.5.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
186
+ "encoder.block.5.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
187
+ "encoder.block.6.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
188
+ "encoder.block.6.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
189
+ "encoder.block.6.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
190
+ "encoder.block.6.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
191
+ "encoder.block.6.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
192
+ "encoder.block.6.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
193
+ "encoder.block.6.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
194
+ "encoder.block.6.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
195
+ "encoder.block.6.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
196
+ "encoder.block.7.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
197
+ "encoder.block.7.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
198
+ "encoder.block.7.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
199
+ "encoder.block.7.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
200
+ "encoder.block.7.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
201
+ "encoder.block.7.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
202
+ "encoder.block.7.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
203
+ "encoder.block.7.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
204
+ "encoder.block.7.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
205
+ "encoder.block.8.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
206
+ "encoder.block.8.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
207
+ "encoder.block.8.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
208
+ "encoder.block.8.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
209
+ "encoder.block.8.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
210
+ "encoder.block.8.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
211
+ "encoder.block.8.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
212
+ "encoder.block.8.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
213
+ "encoder.block.8.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
214
+ "encoder.block.9.layer.0.SelfAttention.k.weight": "model-00001-of-00002.safetensors",
215
+ "encoder.block.9.layer.0.SelfAttention.o.weight": "model-00001-of-00002.safetensors",
216
+ "encoder.block.9.layer.0.SelfAttention.q.weight": "model-00001-of-00002.safetensors",
217
+ "encoder.block.9.layer.0.SelfAttention.v.weight": "model-00001-of-00002.safetensors",
218
+ "encoder.block.9.layer.0.layer_norm.weight": "model-00001-of-00002.safetensors",
219
+ "encoder.block.9.layer.1.DenseReluDense.wi_0.weight": "model-00001-of-00002.safetensors",
220
+ "encoder.block.9.layer.1.DenseReluDense.wi_1.weight": "model-00001-of-00002.safetensors",
221
+ "encoder.block.9.layer.1.DenseReluDense.wo.weight": "model-00001-of-00002.safetensors",
222
+ "encoder.block.9.layer.1.layer_norm.weight": "model-00001-of-00002.safetensors",
223
+ "encoder.final_layer_norm.weight": "model-00002-of-00002.safetensors",
224
+ "shared.weight": "model-00001-of-00002.safetensors"
225
+ }
226
+ }
tokenizer/added_tokens.json ADDED
@@ -0,0 +1,102 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "<extra_id_0>": 32099,
3
+ "<extra_id_10>": 32089,
4
+ "<extra_id_11>": 32088,
5
+ "<extra_id_12>": 32087,
6
+ "<extra_id_13>": 32086,
7
+ "<extra_id_14>": 32085,
8
+ "<extra_id_15>": 32084,
9
+ "<extra_id_16>": 32083,
10
+ "<extra_id_17>": 32082,
11
+ "<extra_id_18>": 32081,
12
+ "<extra_id_19>": 32080,
13
+ "<extra_id_1>": 32098,
14
+ "<extra_id_20>": 32079,
15
+ "<extra_id_21>": 32078,
16
+ "<extra_id_22>": 32077,
17
+ "<extra_id_23>": 32076,
18
+ "<extra_id_24>": 32075,
19
+ "<extra_id_25>": 32074,
20
+ "<extra_id_26>": 32073,
21
+ "<extra_id_27>": 32072,
22
+ "<extra_id_28>": 32071,
23
+ "<extra_id_29>": 32070,
24
+ "<extra_id_2>": 32097,
25
+ "<extra_id_30>": 32069,
26
+ "<extra_id_31>": 32068,
27
+ "<extra_id_32>": 32067,
28
+ "<extra_id_33>": 32066,
29
+ "<extra_id_34>": 32065,
30
+ "<extra_id_35>": 32064,
31
+ "<extra_id_36>": 32063,
32
+ "<extra_id_37>": 32062,
33
+ "<extra_id_38>": 32061,
34
+ "<extra_id_39>": 32060,
35
+ "<extra_id_3>": 32096,
36
+ "<extra_id_40>": 32059,
37
+ "<extra_id_41>": 32058,
38
+ "<extra_id_42>": 32057,
39
+ "<extra_id_43>": 32056,
40
+ "<extra_id_44>": 32055,
41
+ "<extra_id_45>": 32054,
42
+ "<extra_id_46>": 32053,
43
+ "<extra_id_47>": 32052,
44
+ "<extra_id_48>": 32051,
45
+ "<extra_id_49>": 32050,
46
+ "<extra_id_4>": 32095,
47
+ "<extra_id_50>": 32049,
48
+ "<extra_id_51>": 32048,
49
+ "<extra_id_52>": 32047,
50
+ "<extra_id_53>": 32046,
51
+ "<extra_id_54>": 32045,
52
+ "<extra_id_55>": 32044,
53
+ "<extra_id_56>": 32043,
54
+ "<extra_id_57>": 32042,
55
+ "<extra_id_58>": 32041,
56
+ "<extra_id_59>": 32040,
57
+ "<extra_id_5>": 32094,
58
+ "<extra_id_60>": 32039,
59
+ "<extra_id_61>": 32038,
60
+ "<extra_id_62>": 32037,
61
+ "<extra_id_63>": 32036,
62
+ "<extra_id_64>": 32035,
63
+ "<extra_id_65>": 32034,
64
+ "<extra_id_66>": 32033,
65
+ "<extra_id_67>": 32032,
66
+ "<extra_id_68>": 32031,
67
+ "<extra_id_69>": 32030,
68
+ "<extra_id_6>": 32093,
69
+ "<extra_id_70>": 32029,
70
+ "<extra_id_71>": 32028,
71
+ "<extra_id_72>": 32027,
72
+ "<extra_id_73>": 32026,
73
+ "<extra_id_74>": 32025,
74
+ "<extra_id_75>": 32024,
75
+ "<extra_id_76>": 32023,
76
+ "<extra_id_77>": 32022,
77
+ "<extra_id_78>": 32021,
78
+ "<extra_id_79>": 32020,
79
+ "<extra_id_7>": 32092,
80
+ "<extra_id_80>": 32019,
81
+ "<extra_id_81>": 32018,
82
+ "<extra_id_82>": 32017,
83
+ "<extra_id_83>": 32016,
84
+ "<extra_id_84>": 32015,
85
+ "<extra_id_85>": 32014,
86
+ "<extra_id_86>": 32013,
87
+ "<extra_id_87>": 32012,
88
+ "<extra_id_88>": 32011,
89
+ "<extra_id_89>": 32010,
90
+ "<extra_id_8>": 32091,
91
+ "<extra_id_90>": 32009,
92
+ "<extra_id_91>": 32008,
93
+ "<extra_id_92>": 32007,
94
+ "<extra_id_93>": 32006,
95
+ "<extra_id_94>": 32005,
96
+ "<extra_id_95>": 32004,
97
+ "<extra_id_96>": 32003,
98
+ "<extra_id_97>": 32002,
99
+ "<extra_id_98>": 32001,
100
+ "<extra_id_99>": 32000,
101
+ "<extra_id_9>": 32090
102
+ }
tokenizer/special_tokens_map.json ADDED
@@ -0,0 +1,125 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<extra_id_0>",
4
+ "<extra_id_1>",
5
+ "<extra_id_2>",
6
+ "<extra_id_3>",
7
+ "<extra_id_4>",
8
+ "<extra_id_5>",
9
+ "<extra_id_6>",
10
+ "<extra_id_7>",
11
+ "<extra_id_8>",
12
+ "<extra_id_9>",
13
+ "<extra_id_10>",
14
+ "<extra_id_11>",
15
+ "<extra_id_12>",
16
+ "<extra_id_13>",
17
+ "<extra_id_14>",
18
+ "<extra_id_15>",
19
+ "<extra_id_16>",
20
+ "<extra_id_17>",
21
+ "<extra_id_18>",
22
+ "<extra_id_19>",
23
+ "<extra_id_20>",
24
+ "<extra_id_21>",
25
+ "<extra_id_22>",
26
+ "<extra_id_23>",
27
+ "<extra_id_24>",
28
+ "<extra_id_25>",
29
+ "<extra_id_26>",
30
+ "<extra_id_27>",
31
+ "<extra_id_28>",
32
+ "<extra_id_29>",
33
+ "<extra_id_30>",
34
+ "<extra_id_31>",
35
+ "<extra_id_32>",
36
+ "<extra_id_33>",
37
+ "<extra_id_34>",
38
+ "<extra_id_35>",
39
+ "<extra_id_36>",
40
+ "<extra_id_37>",
41
+ "<extra_id_38>",
42
+ "<extra_id_39>",
43
+ "<extra_id_40>",
44
+ "<extra_id_41>",
45
+ "<extra_id_42>",
46
+ "<extra_id_43>",
47
+ "<extra_id_44>",
48
+ "<extra_id_45>",
49
+ "<extra_id_46>",
50
+ "<extra_id_47>",
51
+ "<extra_id_48>",
52
+ "<extra_id_49>",
53
+ "<extra_id_50>",
54
+ "<extra_id_51>",
55
+ "<extra_id_52>",
56
+ "<extra_id_53>",
57
+ "<extra_id_54>",
58
+ "<extra_id_55>",
59
+ "<extra_id_56>",
60
+ "<extra_id_57>",
61
+ "<extra_id_58>",
62
+ "<extra_id_59>",
63
+ "<extra_id_60>",
64
+ "<extra_id_61>",
65
+ "<extra_id_62>",
66
+ "<extra_id_63>",
67
+ "<extra_id_64>",
68
+ "<extra_id_65>",
69
+ "<extra_id_66>",
70
+ "<extra_id_67>",
71
+ "<extra_id_68>",
72
+ "<extra_id_69>",
73
+ "<extra_id_70>",
74
+ "<extra_id_71>",
75
+ "<extra_id_72>",
76
+ "<extra_id_73>",
77
+ "<extra_id_74>",
78
+ "<extra_id_75>",
79
+ "<extra_id_76>",
80
+ "<extra_id_77>",
81
+ "<extra_id_78>",
82
+ "<extra_id_79>",
83
+ "<extra_id_80>",
84
+ "<extra_id_81>",
85
+ "<extra_id_82>",
86
+ "<extra_id_83>",
87
+ "<extra_id_84>",
88
+ "<extra_id_85>",
89
+ "<extra_id_86>",
90
+ "<extra_id_87>",
91
+ "<extra_id_88>",
92
+ "<extra_id_89>",
93
+ "<extra_id_90>",
94
+ "<extra_id_91>",
95
+ "<extra_id_92>",
96
+ "<extra_id_93>",
97
+ "<extra_id_94>",
98
+ "<extra_id_95>",
99
+ "<extra_id_96>",
100
+ "<extra_id_97>",
101
+ "<extra_id_98>",
102
+ "<extra_id_99>"
103
+ ],
104
+ "eos_token": {
105
+ "content": "</s>",
106
+ "lstrip": false,
107
+ "normalized": false,
108
+ "rstrip": false,
109
+ "single_word": false
110
+ },
111
+ "pad_token": {
112
+ "content": "<pad>",
113
+ "lstrip": false,
114
+ "normalized": false,
115
+ "rstrip": false,
116
+ "single_word": false
117
+ },
118
+ "unk_token": {
119
+ "content": "<unk>",
120
+ "lstrip": false,
121
+ "normalized": false,
122
+ "rstrip": false,
123
+ "single_word": false
124
+ }
125
+ }
tokenizer/spiece.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d60acb128cf7b7f2536e8f38a5b18a05535c9e14c7a355904270e15b0945ea86
3
+ size 791656
tokenizer/tokenizer_config.json ADDED
@@ -0,0 +1,940 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": true,
3
+ "added_tokens_decoder": {
4
+ "0": {
5
+ "content": "<pad>",
6
+ "lstrip": false,
7
+ "normalized": false,
8
+ "rstrip": false,
9
+ "single_word": false,
10
+ "special": true
11
+ },
12
+ "1": {
13
+ "content": "</s>",
14
+ "lstrip": false,
15
+ "normalized": false,
16
+ "rstrip": false,
17
+ "single_word": false,
18
+ "special": true
19
+ },
20
+ "2": {
21
+ "content": "<unk>",
22
+ "lstrip": false,
23
+ "normalized": false,
24
+ "rstrip": false,
25
+ "single_word": false,
26
+ "special": true
27
+ },
28
+ "32000": {
29
+ "content": "<extra_id_99>",
30
+ "lstrip": true,
31
+ "normalized": false,
32
+ "rstrip": true,
33
+ "single_word": false,
34
+ "special": true
35
+ },
36
+ "32001": {
37
+ "content": "<extra_id_98>",
38
+ "lstrip": true,
39
+ "normalized": false,
40
+ "rstrip": true,
41
+ "single_word": false,
42
+ "special": true
43
+ },
44
+ "32002": {
45
+ "content": "<extra_id_97>",
46
+ "lstrip": true,
47
+ "normalized": false,
48
+ "rstrip": true,
49
+ "single_word": false,
50
+ "special": true
51
+ },
52
+ "32003": {
53
+ "content": "<extra_id_96>",
54
+ "lstrip": true,
55
+ "normalized": false,
56
+ "rstrip": true,
57
+ "single_word": false,
58
+ "special": true
59
+ },
60
+ "32004": {
61
+ "content": "<extra_id_95>",
62
+ "lstrip": true,
63
+ "normalized": false,
64
+ "rstrip": true,
65
+ "single_word": false,
66
+ "special": true
67
+ },
68
+ "32005": {
69
+ "content": "<extra_id_94>",
70
+ "lstrip": true,
71
+ "normalized": false,
72
+ "rstrip": true,
73
+ "single_word": false,
74
+ "special": true
75
+ },
76
+ "32006": {
77
+ "content": "<extra_id_93>",
78
+ "lstrip": true,
79
+ "normalized": false,
80
+ "rstrip": true,
81
+ "single_word": false,
82
+ "special": true
83
+ },
84
+ "32007": {
85
+ "content": "<extra_id_92>",
86
+ "lstrip": true,
87
+ "normalized": false,
88
+ "rstrip": true,
89
+ "single_word": false,
90
+ "special": true
91
+ },
92
+ "32008": {
93
+ "content": "<extra_id_91>",
94
+ "lstrip": true,
95
+ "normalized": false,
96
+ "rstrip": true,
97
+ "single_word": false,
98
+ "special": true
99
+ },
100
+ "32009": {
101
+ "content": "<extra_id_90>",
102
+ "lstrip": true,
103
+ "normalized": false,
104
+ "rstrip": true,
105
+ "single_word": false,
106
+ "special": true
107
+ },
108
+ "32010": {
109
+ "content": "<extra_id_89>",
110
+ "lstrip": true,
111
+ "normalized": false,
112
+ "rstrip": true,
113
+ "single_word": false,
114
+ "special": true
115
+ },
116
+ "32011": {
117
+ "content": "<extra_id_88>",
118
+ "lstrip": true,
119
+ "normalized": false,
120
+ "rstrip": true,
121
+ "single_word": false,
122
+ "special": true
123
+ },
124
+ "32012": {
125
+ "content": "<extra_id_87>",
126
+ "lstrip": true,
127
+ "normalized": false,
128
+ "rstrip": true,
129
+ "single_word": false,
130
+ "special": true
131
+ },
132
+ "32013": {
133
+ "content": "<extra_id_86>",
134
+ "lstrip": true,
135
+ "normalized": false,
136
+ "rstrip": true,
137
+ "single_word": false,
138
+ "special": true
139
+ },
140
+ "32014": {
141
+ "content": "<extra_id_85>",
142
+ "lstrip": true,
143
+ "normalized": false,
144
+ "rstrip": true,
145
+ "single_word": false,
146
+ "special": true
147
+ },
148
+ "32015": {
149
+ "content": "<extra_id_84>",
150
+ "lstrip": true,
151
+ "normalized": false,
152
+ "rstrip": true,
153
+ "single_word": false,
154
+ "special": true
155
+ },
156
+ "32016": {
157
+ "content": "<extra_id_83>",
158
+ "lstrip": true,
159
+ "normalized": false,
160
+ "rstrip": true,
161
+ "single_word": false,
162
+ "special": true
163
+ },
164
+ "32017": {
165
+ "content": "<extra_id_82>",
166
+ "lstrip": true,
167
+ "normalized": false,
168
+ "rstrip": true,
169
+ "single_word": false,
170
+ "special": true
171
+ },
172
+ "32018": {
173
+ "content": "<extra_id_81>",
174
+ "lstrip": true,
175
+ "normalized": false,
176
+ "rstrip": true,
177
+ "single_word": false,
178
+ "special": true
179
+ },
180
+ "32019": {
181
+ "content": "<extra_id_80>",
182
+ "lstrip": true,
183
+ "normalized": false,
184
+ "rstrip": true,
185
+ "single_word": false,
186
+ "special": true
187
+ },
188
+ "32020": {
189
+ "content": "<extra_id_79>",
190
+ "lstrip": true,
191
+ "normalized": false,
192
+ "rstrip": true,
193
+ "single_word": false,
194
+ "special": true
195
+ },
196
+ "32021": {
197
+ "content": "<extra_id_78>",
198
+ "lstrip": true,
199
+ "normalized": false,
200
+ "rstrip": true,
201
+ "single_word": false,
202
+ "special": true
203
+ },
204
+ "32022": {
205
+ "content": "<extra_id_77>",
206
+ "lstrip": true,
207
+ "normalized": false,
208
+ "rstrip": true,
209
+ "single_word": false,
210
+ "special": true
211
+ },
212
+ "32023": {
213
+ "content": "<extra_id_76>",
214
+ "lstrip": true,
215
+ "normalized": false,
216
+ "rstrip": true,
217
+ "single_word": false,
218
+ "special": true
219
+ },
220
+ "32024": {
221
+ "content": "<extra_id_75>",
222
+ "lstrip": true,
223
+ "normalized": false,
224
+ "rstrip": true,
225
+ "single_word": false,
226
+ "special": true
227
+ },
228
+ "32025": {
229
+ "content": "<extra_id_74>",
230
+ "lstrip": true,
231
+ "normalized": false,
232
+ "rstrip": true,
233
+ "single_word": false,
234
+ "special": true
235
+ },
236
+ "32026": {
237
+ "content": "<extra_id_73>",
238
+ "lstrip": true,
239
+ "normalized": false,
240
+ "rstrip": true,
241
+ "single_word": false,
242
+ "special": true
243
+ },
244
+ "32027": {
245
+ "content": "<extra_id_72>",
246
+ "lstrip": true,
247
+ "normalized": false,
248
+ "rstrip": true,
249
+ "single_word": false,
250
+ "special": true
251
+ },
252
+ "32028": {
253
+ "content": "<extra_id_71>",
254
+ "lstrip": true,
255
+ "normalized": false,
256
+ "rstrip": true,
257
+ "single_word": false,
258
+ "special": true
259
+ },
260
+ "32029": {
261
+ "content": "<extra_id_70>",
262
+ "lstrip": true,
263
+ "normalized": false,
264
+ "rstrip": true,
265
+ "single_word": false,
266
+ "special": true
267
+ },
268
+ "32030": {
269
+ "content": "<extra_id_69>",
270
+ "lstrip": true,
271
+ "normalized": false,
272
+ "rstrip": true,
273
+ "single_word": false,
274
+ "special": true
275
+ },
276
+ "32031": {
277
+ "content": "<extra_id_68>",
278
+ "lstrip": true,
279
+ "normalized": false,
280
+ "rstrip": true,
281
+ "single_word": false,
282
+ "special": true
283
+ },
284
+ "32032": {
285
+ "content": "<extra_id_67>",
286
+ "lstrip": true,
287
+ "normalized": false,
288
+ "rstrip": true,
289
+ "single_word": false,
290
+ "special": true
291
+ },
292
+ "32033": {
293
+ "content": "<extra_id_66>",
294
+ "lstrip": true,
295
+ "normalized": false,
296
+ "rstrip": true,
297
+ "single_word": false,
298
+ "special": true
299
+ },
300
+ "32034": {
301
+ "content": "<extra_id_65>",
302
+ "lstrip": true,
303
+ "normalized": false,
304
+ "rstrip": true,
305
+ "single_word": false,
306
+ "special": true
307
+ },
308
+ "32035": {
309
+ "content": "<extra_id_64>",
310
+ "lstrip": true,
311
+ "normalized": false,
312
+ "rstrip": true,
313
+ "single_word": false,
314
+ "special": true
315
+ },
316
+ "32036": {
317
+ "content": "<extra_id_63>",
318
+ "lstrip": true,
319
+ "normalized": false,
320
+ "rstrip": true,
321
+ "single_word": false,
322
+ "special": true
323
+ },
324
+ "32037": {
325
+ "content": "<extra_id_62>",
326
+ "lstrip": true,
327
+ "normalized": false,
328
+ "rstrip": true,
329
+ "single_word": false,
330
+ "special": true
331
+ },
332
+ "32038": {
333
+ "content": "<extra_id_61>",
334
+ "lstrip": true,
335
+ "normalized": false,
336
+ "rstrip": true,
337
+ "single_word": false,
338
+ "special": true
339
+ },
340
+ "32039": {
341
+ "content": "<extra_id_60>",
342
+ "lstrip": true,
343
+ "normalized": false,
344
+ "rstrip": true,
345
+ "single_word": false,
346
+ "special": true
347
+ },
348
+ "32040": {
349
+ "content": "<extra_id_59>",
350
+ "lstrip": true,
351
+ "normalized": false,
352
+ "rstrip": true,
353
+ "single_word": false,
354
+ "special": true
355
+ },
356
+ "32041": {
357
+ "content": "<extra_id_58>",
358
+ "lstrip": true,
359
+ "normalized": false,
360
+ "rstrip": true,
361
+ "single_word": false,
362
+ "special": true
363
+ },
364
+ "32042": {
365
+ "content": "<extra_id_57>",
366
+ "lstrip": true,
367
+ "normalized": false,
368
+ "rstrip": true,
369
+ "single_word": false,
370
+ "special": true
371
+ },
372
+ "32043": {
373
+ "content": "<extra_id_56>",
374
+ "lstrip": true,
375
+ "normalized": false,
376
+ "rstrip": true,
377
+ "single_word": false,
378
+ "special": true
379
+ },
380
+ "32044": {
381
+ "content": "<extra_id_55>",
382
+ "lstrip": true,
383
+ "normalized": false,
384
+ "rstrip": true,
385
+ "single_word": false,
386
+ "special": true
387
+ },
388
+ "32045": {
389
+ "content": "<extra_id_54>",
390
+ "lstrip": true,
391
+ "normalized": false,
392
+ "rstrip": true,
393
+ "single_word": false,
394
+ "special": true
395
+ },
396
+ "32046": {
397
+ "content": "<extra_id_53>",
398
+ "lstrip": true,
399
+ "normalized": false,
400
+ "rstrip": true,
401
+ "single_word": false,
402
+ "special": true
403
+ },
404
+ "32047": {
405
+ "content": "<extra_id_52>",
406
+ "lstrip": true,
407
+ "normalized": false,
408
+ "rstrip": true,
409
+ "single_word": false,
410
+ "special": true
411
+ },
412
+ "32048": {
413
+ "content": "<extra_id_51>",
414
+ "lstrip": true,
415
+ "normalized": false,
416
+ "rstrip": true,
417
+ "single_word": false,
418
+ "special": true
419
+ },
420
+ "32049": {
421
+ "content": "<extra_id_50>",
422
+ "lstrip": true,
423
+ "normalized": false,
424
+ "rstrip": true,
425
+ "single_word": false,
426
+ "special": true
427
+ },
428
+ "32050": {
429
+ "content": "<extra_id_49>",
430
+ "lstrip": true,
431
+ "normalized": false,
432
+ "rstrip": true,
433
+ "single_word": false,
434
+ "special": true
435
+ },
436
+ "32051": {
437
+ "content": "<extra_id_48>",
438
+ "lstrip": true,
439
+ "normalized": false,
440
+ "rstrip": true,
441
+ "single_word": false,
442
+ "special": true
443
+ },
444
+ "32052": {
445
+ "content": "<extra_id_47>",
446
+ "lstrip": true,
447
+ "normalized": false,
448
+ "rstrip": true,
449
+ "single_word": false,
450
+ "special": true
451
+ },
452
+ "32053": {
453
+ "content": "<extra_id_46>",
454
+ "lstrip": true,
455
+ "normalized": false,
456
+ "rstrip": true,
457
+ "single_word": false,
458
+ "special": true
459
+ },
460
+ "32054": {
461
+ "content": "<extra_id_45>",
462
+ "lstrip": true,
463
+ "normalized": false,
464
+ "rstrip": true,
465
+ "single_word": false,
466
+ "special": true
467
+ },
468
+ "32055": {
469
+ "content": "<extra_id_44>",
470
+ "lstrip": true,
471
+ "normalized": false,
472
+ "rstrip": true,
473
+ "single_word": false,
474
+ "special": true
475
+ },
476
+ "32056": {
477
+ "content": "<extra_id_43>",
478
+ "lstrip": true,
479
+ "normalized": false,
480
+ "rstrip": true,
481
+ "single_word": false,
482
+ "special": true
483
+ },
484
+ "32057": {
485
+ "content": "<extra_id_42>",
486
+ "lstrip": true,
487
+ "normalized": false,
488
+ "rstrip": true,
489
+ "single_word": false,
490
+ "special": true
491
+ },
492
+ "32058": {
493
+ "content": "<extra_id_41>",
494
+ "lstrip": true,
495
+ "normalized": false,
496
+ "rstrip": true,
497
+ "single_word": false,
498
+ "special": true
499
+ },
500
+ "32059": {
501
+ "content": "<extra_id_40>",
502
+ "lstrip": true,
503
+ "normalized": false,
504
+ "rstrip": true,
505
+ "single_word": false,
506
+ "special": true
507
+ },
508
+ "32060": {
509
+ "content": "<extra_id_39>",
510
+ "lstrip": true,
511
+ "normalized": false,
512
+ "rstrip": true,
513
+ "single_word": false,
514
+ "special": true
515
+ },
516
+ "32061": {
517
+ "content": "<extra_id_38>",
518
+ "lstrip": true,
519
+ "normalized": false,
520
+ "rstrip": true,
521
+ "single_word": false,
522
+ "special": true
523
+ },
524
+ "32062": {
525
+ "content": "<extra_id_37>",
526
+ "lstrip": true,
527
+ "normalized": false,
528
+ "rstrip": true,
529
+ "single_word": false,
530
+ "special": true
531
+ },
532
+ "32063": {
533
+ "content": "<extra_id_36>",
534
+ "lstrip": true,
535
+ "normalized": false,
536
+ "rstrip": true,
537
+ "single_word": false,
538
+ "special": true
539
+ },
540
+ "32064": {
541
+ "content": "<extra_id_35>",
542
+ "lstrip": true,
543
+ "normalized": false,
544
+ "rstrip": true,
545
+ "single_word": false,
546
+ "special": true
547
+ },
548
+ "32065": {
549
+ "content": "<extra_id_34>",
550
+ "lstrip": true,
551
+ "normalized": false,
552
+ "rstrip": true,
553
+ "single_word": false,
554
+ "special": true
555
+ },
556
+ "32066": {
557
+ "content": "<extra_id_33>",
558
+ "lstrip": true,
559
+ "normalized": false,
560
+ "rstrip": true,
561
+ "single_word": false,
562
+ "special": true
563
+ },
564
+ "32067": {
565
+ "content": "<extra_id_32>",
566
+ "lstrip": true,
567
+ "normalized": false,
568
+ "rstrip": true,
569
+ "single_word": false,
570
+ "special": true
571
+ },
572
+ "32068": {
573
+ "content": "<extra_id_31>",
574
+ "lstrip": true,
575
+ "normalized": false,
576
+ "rstrip": true,
577
+ "single_word": false,
578
+ "special": true
579
+ },
580
+ "32069": {
581
+ "content": "<extra_id_30>",
582
+ "lstrip": true,
583
+ "normalized": false,
584
+ "rstrip": true,
585
+ "single_word": false,
586
+ "special": true
587
+ },
588
+ "32070": {
589
+ "content": "<extra_id_29>",
590
+ "lstrip": true,
591
+ "normalized": false,
592
+ "rstrip": true,
593
+ "single_word": false,
594
+ "special": true
595
+ },
596
+ "32071": {
597
+ "content": "<extra_id_28>",
598
+ "lstrip": true,
599
+ "normalized": false,
600
+ "rstrip": true,
601
+ "single_word": false,
602
+ "special": true
603
+ },
604
+ "32072": {
605
+ "content": "<extra_id_27>",
606
+ "lstrip": true,
607
+ "normalized": false,
608
+ "rstrip": true,
609
+ "single_word": false,
610
+ "special": true
611
+ },
612
+ "32073": {
613
+ "content": "<extra_id_26>",
614
+ "lstrip": true,
615
+ "normalized": false,
616
+ "rstrip": true,
617
+ "single_word": false,
618
+ "special": true
619
+ },
620
+ "32074": {
621
+ "content": "<extra_id_25>",
622
+ "lstrip": true,
623
+ "normalized": false,
624
+ "rstrip": true,
625
+ "single_word": false,
626
+ "special": true
627
+ },
628
+ "32075": {
629
+ "content": "<extra_id_24>",
630
+ "lstrip": true,
631
+ "normalized": false,
632
+ "rstrip": true,
633
+ "single_word": false,
634
+ "special": true
635
+ },
636
+ "32076": {
637
+ "content": "<extra_id_23>",
638
+ "lstrip": true,
639
+ "normalized": false,
640
+ "rstrip": true,
641
+ "single_word": false,
642
+ "special": true
643
+ },
644
+ "32077": {
645
+ "content": "<extra_id_22>",
646
+ "lstrip": true,
647
+ "normalized": false,
648
+ "rstrip": true,
649
+ "single_word": false,
650
+ "special": true
651
+ },
652
+ "32078": {
653
+ "content": "<extra_id_21>",
654
+ "lstrip": true,
655
+ "normalized": false,
656
+ "rstrip": true,
657
+ "single_word": false,
658
+ "special": true
659
+ },
660
+ "32079": {
661
+ "content": "<extra_id_20>",
662
+ "lstrip": true,
663
+ "normalized": false,
664
+ "rstrip": true,
665
+ "single_word": false,
666
+ "special": true
667
+ },
668
+ "32080": {
669
+ "content": "<extra_id_19>",
670
+ "lstrip": true,
671
+ "normalized": false,
672
+ "rstrip": true,
673
+ "single_word": false,
674
+ "special": true
675
+ },
676
+ "32081": {
677
+ "content": "<extra_id_18>",
678
+ "lstrip": true,
679
+ "normalized": false,
680
+ "rstrip": true,
681
+ "single_word": false,
682
+ "special": true
683
+ },
684
+ "32082": {
685
+ "content": "<extra_id_17>",
686
+ "lstrip": true,
687
+ "normalized": false,
688
+ "rstrip": true,
689
+ "single_word": false,
690
+ "special": true
691
+ },
692
+ "32083": {
693
+ "content": "<extra_id_16>",
694
+ "lstrip": true,
695
+ "normalized": false,
696
+ "rstrip": true,
697
+ "single_word": false,
698
+ "special": true
699
+ },
700
+ "32084": {
701
+ "content": "<extra_id_15>",
702
+ "lstrip": true,
703
+ "normalized": false,
704
+ "rstrip": true,
705
+ "single_word": false,
706
+ "special": true
707
+ },
708
+ "32085": {
709
+ "content": "<extra_id_14>",
710
+ "lstrip": true,
711
+ "normalized": false,
712
+ "rstrip": true,
713
+ "single_word": false,
714
+ "special": true
715
+ },
716
+ "32086": {
717
+ "content": "<extra_id_13>",
718
+ "lstrip": true,
719
+ "normalized": false,
720
+ "rstrip": true,
721
+ "single_word": false,
722
+ "special": true
723
+ },
724
+ "32087": {
725
+ "content": "<extra_id_12>",
726
+ "lstrip": true,
727
+ "normalized": false,
728
+ "rstrip": true,
729
+ "single_word": false,
730
+ "special": true
731
+ },
732
+ "32088": {
733
+ "content": "<extra_id_11>",
734
+ "lstrip": true,
735
+ "normalized": false,
736
+ "rstrip": true,
737
+ "single_word": false,
738
+ "special": true
739
+ },
740
+ "32089": {
741
+ "content": "<extra_id_10>",
742
+ "lstrip": true,
743
+ "normalized": false,
744
+ "rstrip": true,
745
+ "single_word": false,
746
+ "special": true
747
+ },
748
+ "32090": {
749
+ "content": "<extra_id_9>",
750
+ "lstrip": true,
751
+ "normalized": false,
752
+ "rstrip": true,
753
+ "single_word": false,
754
+ "special": true
755
+ },
756
+ "32091": {
757
+ "content": "<extra_id_8>",
758
+ "lstrip": true,
759
+ "normalized": false,
760
+ "rstrip": true,
761
+ "single_word": false,
762
+ "special": true
763
+ },
764
+ "32092": {
765
+ "content": "<extra_id_7>",
766
+ "lstrip": true,
767
+ "normalized": false,
768
+ "rstrip": true,
769
+ "single_word": false,
770
+ "special": true
771
+ },
772
+ "32093": {
773
+ "content": "<extra_id_6>",
774
+ "lstrip": true,
775
+ "normalized": false,
776
+ "rstrip": true,
777
+ "single_word": false,
778
+ "special": true
779
+ },
780
+ "32094": {
781
+ "content": "<extra_id_5>",
782
+ "lstrip": true,
783
+ "normalized": false,
784
+ "rstrip": true,
785
+ "single_word": false,
786
+ "special": true
787
+ },
788
+ "32095": {
789
+ "content": "<extra_id_4>",
790
+ "lstrip": true,
791
+ "normalized": false,
792
+ "rstrip": true,
793
+ "single_word": false,
794
+ "special": true
795
+ },
796
+ "32096": {
797
+ "content": "<extra_id_3>",
798
+ "lstrip": true,
799
+ "normalized": false,
800
+ "rstrip": true,
801
+ "single_word": false,
802
+ "special": true
803
+ },
804
+ "32097": {
805
+ "content": "<extra_id_2>",
806
+ "lstrip": true,
807
+ "normalized": false,
808
+ "rstrip": true,
809
+ "single_word": false,
810
+ "special": true
811
+ },
812
+ "32098": {
813
+ "content": "<extra_id_1>",
814
+ "lstrip": true,
815
+ "normalized": false,
816
+ "rstrip": true,
817
+ "single_word": false,
818
+ "special": true
819
+ },
820
+ "32099": {
821
+ "content": "<extra_id_0>",
822
+ "lstrip": true,
823
+ "normalized": false,
824
+ "rstrip": true,
825
+ "single_word": false,
826
+ "special": true
827
+ }
828
+ },
829
+ "additional_special_tokens": [
830
+ "<extra_id_0>",
831
+ "<extra_id_1>",
832
+ "<extra_id_2>",
833
+ "<extra_id_3>",
834
+ "<extra_id_4>",
835
+ "<extra_id_5>",
836
+ "<extra_id_6>",
837
+ "<extra_id_7>",
838
+ "<extra_id_8>",
839
+ "<extra_id_9>",
840
+ "<extra_id_10>",
841
+ "<extra_id_11>",
842
+ "<extra_id_12>",
843
+ "<extra_id_13>",
844
+ "<extra_id_14>",
845
+ "<extra_id_15>",
846
+ "<extra_id_16>",
847
+ "<extra_id_17>",
848
+ "<extra_id_18>",
849
+ "<extra_id_19>",
850
+ "<extra_id_20>",
851
+ "<extra_id_21>",
852
+ "<extra_id_22>",
853
+ "<extra_id_23>",
854
+ "<extra_id_24>",
855
+ "<extra_id_25>",
856
+ "<extra_id_26>",
857
+ "<extra_id_27>",
858
+ "<extra_id_28>",
859
+ "<extra_id_29>",
860
+ "<extra_id_30>",
861
+ "<extra_id_31>",
862
+ "<extra_id_32>",
863
+ "<extra_id_33>",
864
+ "<extra_id_34>",
865
+ "<extra_id_35>",
866
+ "<extra_id_36>",
867
+ "<extra_id_37>",
868
+ "<extra_id_38>",
869
+ "<extra_id_39>",
870
+ "<extra_id_40>",
871
+ "<extra_id_41>",
872
+ "<extra_id_42>",
873
+ "<extra_id_43>",
874
+ "<extra_id_44>",
875
+ "<extra_id_45>",
876
+ "<extra_id_46>",
877
+ "<extra_id_47>",
878
+ "<extra_id_48>",
879
+ "<extra_id_49>",
880
+ "<extra_id_50>",
881
+ "<extra_id_51>",
882
+ "<extra_id_52>",
883
+ "<extra_id_53>",
884
+ "<extra_id_54>",
885
+ "<extra_id_55>",
886
+ "<extra_id_56>",
887
+ "<extra_id_57>",
888
+ "<extra_id_58>",
889
+ "<extra_id_59>",
890
+ "<extra_id_60>",
891
+ "<extra_id_61>",
892
+ "<extra_id_62>",
893
+ "<extra_id_63>",
894
+ "<extra_id_64>",
895
+ "<extra_id_65>",
896
+ "<extra_id_66>",
897
+ "<extra_id_67>",
898
+ "<extra_id_68>",
899
+ "<extra_id_69>",
900
+ "<extra_id_70>",
901
+ "<extra_id_71>",
902
+ "<extra_id_72>",
903
+ "<extra_id_73>",
904
+ "<extra_id_74>",
905
+ "<extra_id_75>",
906
+ "<extra_id_76>",
907
+ "<extra_id_77>",
908
+ "<extra_id_78>",
909
+ "<extra_id_79>",
910
+ "<extra_id_80>",
911
+ "<extra_id_81>",
912
+ "<extra_id_82>",
913
+ "<extra_id_83>",
914
+ "<extra_id_84>",
915
+ "<extra_id_85>",
916
+ "<extra_id_86>",
917
+ "<extra_id_87>",
918
+ "<extra_id_88>",
919
+ "<extra_id_89>",
920
+ "<extra_id_90>",
921
+ "<extra_id_91>",
922
+ "<extra_id_92>",
923
+ "<extra_id_93>",
924
+ "<extra_id_94>",
925
+ "<extra_id_95>",
926
+ "<extra_id_96>",
927
+ "<extra_id_97>",
928
+ "<extra_id_98>",
929
+ "<extra_id_99>"
930
+ ],
931
+ "clean_up_tokenization_spaces": true,
932
+ "eos_token": "</s>",
933
+ "extra_ids": 100,
934
+ "legacy": true,
935
+ "model_max_length": 226,
936
+ "pad_token": "<pad>",
937
+ "sp_model_kwargs": {},
938
+ "tokenizer_class": "T5Tokenizer",
939
+ "unk_token": "<unk>"
940
+ }
transformer/config.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_class_name": "CogVideoXTransformer3DModel",
3
+ "_diffusers_version": "0.30.0.dev0",
4
+ "activation_fn": "gelu-approximate",
5
+ "attention_bias": true,
6
+ "attention_head_dim": 64,
7
+ "dropout": 0.0,
8
+ "flip_sin_to_cos": true,
9
+ "freq_shift": 0,
10
+ "in_channels": 16,
11
+ "max_text_seq_length": 226,
12
+ "norm_elementwise_affine": true,
13
+ "norm_eps": 1e-05,
14
+ "num_attention_heads": 30,
15
+ "num_layers": 30,
16
+ "out_channels": 16,
17
+ "patch_size": 2,
18
+ "sample_frames": 49,
19
+ "sample_height": 60,
20
+ "sample_width": 90,
21
+ "spatial_interpolation_scale": 1.875,
22
+ "temporal_compression_ratio": 4,
23
+ "temporal_interpolation_scale": 1.0,
24
+ "text_embed_dim": 4096,
25
+ "time_embed_dim": 512,
26
+ "timestep_activation_fn": "silu",
27
+ "use_rotary_positional_embeddings": false
28
+ }
transformer/diffusion_pytorch_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8fbb6a5e67c70885a8ed8e33df144ac61253e45977be5035fa18cfdf77d386c7
3
+ size 3387650264
vae/config.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_class_name": "AutoencoderKLCogVideoX",
3
+ "_diffusers_version": "0.30.0.dev0",
4
+ "act_fn": "silu",
5
+ "block_out_channels": [
6
+ 128,
7
+ 256,
8
+ 256,
9
+ 512
10
+ ],
11
+ "down_block_types": [
12
+ "CogVideoXDownBlock3D",
13
+ "CogVideoXDownBlock3D",
14
+ "CogVideoXDownBlock3D",
15
+ "CogVideoXDownBlock3D"
16
+ ],
17
+ "force_upcast": true,
18
+ "in_channels": 3,
19
+ "latent_channels": 16,
20
+ "latents_mean": null,
21
+ "latents_std": null,
22
+ "layers_per_block": 3,
23
+ "norm_eps": 1e-06,
24
+ "norm_num_groups": 32,
25
+ "out_channels": 3,
26
+ "sample_height": 480,
27
+ "sample_width": 720,
28
+ "scaling_factor": 1.15258426,
29
+ "shift_factor": null,
30
+ "temporal_compression_ratio": 4,
31
+ "up_block_types": [
32
+ "CogVideoXUpBlock3D",
33
+ "CogVideoXUpBlock3D",
34
+ "CogVideoXUpBlock3D",
35
+ "CogVideoXUpBlock3D"
36
+ ],
37
+ "use_post_quant_conv": false,
38
+ "use_quant_conv": false
39
+ }
vae/diffusion_pytorch_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e25e94a8fc70774349bb4a03b8ef272f5d80f934863f7b0552c37c6a74f91542
3
+ size 431220702