xinyun/ComfyUI-CogVideoXWrapper
1
0
Fork 0
mirror of https://git.datalinker.icu/kijai/ComfyUI-CogVideoXWrapper.git synced 2025-12-08 20:34:23 +08:00
Code Issues Packages Projects Releases Wiki Activity

update examples, fix vid2vid (2b)

Browse Source
This commit is contained in:
kijai 2024-08-27 18:32:57 +03:00
parent a30037feb1
commit 46fe99f102
4 changed files with 326 additions and 882 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

113
examples/cogvideo_long_01.json → examples/cogvideo_2b_long_02.json
View File

@ -30,8 +30,8 @@
"links": [
55
],
"shape": 3,
"slot_index": 0
"slot_index": 0,
"shape": 3
}
],
"properties": {
@ -63,8 +63,8 @@
54,
56
],
"shape": 3,
"slot_index": 0
"slot_index": 0,
"shape": 3
}
],
"properties": {
@ -103,8 +103,8 @@
"links": [
57
],
"shape": 3,
"slot_index": 0
"slot_index": 0,
"shape": 3
}
],
"properties": {
@ -114,38 +114,6 @@
""
]
},
{
"id": 1,
"type": "DownloadAndLoadCogVideoModel",
"pos": [
649,
182
],
"size": {
"0": 315,
"1": 58
},
"flags": {},
"order": 1,
"mode": 0,
"outputs": [
{
"name": "cogvideo_pipe",
"type": "COGVIDEOPIPE",
"links": [
36
],
"shape": 3,
"slot_index": 0
}
],
"properties": {
"Node name for S&R": "DownloadAndLoadCogVideoModel"
},
"widgets_values": [
"bf16"
]
},
{
"id": 11,
"type": "CogVideoDecode",
@ -155,7 +123,7 @@
],
"size": {
"0": 210,
"1": 46
"1": 78
},
"flags": {},
"order": 5,
@ -179,13 +147,16 @@
"links": [
59
],
"shape": 3,
"slot_index": 0
"slot_index": 0,
"shape": 3
}
],
"properties": {
"Node name for S&R": "CogVideoDecode"
}
},
"widgets_values": [
false
]
},
{
"id": 33,
@ -215,8 +186,8 @@
"links": [
60
],
"shape": 3,
"slot_index": 0
"slot_index": 0,
"shape": 3
},
{
"name": "720 width",
@ -231,7 +202,7 @@
"shape": 3
},
{
"name": "122 count",
"name": "32 count",
"type": "INT",
"links": null,
"shape": 3
@ -302,18 +273,50 @@
"widgets_values": [
480,
720,
128,
8,
96,
25,
6,
6,
806286757407563,
"fixed",
"DDIM",
48,
12,
8,
1
]
},
{
"id": 1,
"type": "DownloadAndLoadCogVideoModel",
"pos": [
649,
182
],
"size": {
"0": 315,
"1": 82
},
"flags": {},
"order": 1,
"mode": 0,
"outputs": [
{
"name": "cogvideo_pipe",
"type": "COGVIDEOPIPE",
"links": [
36
],
"slot_index": 0,
"shape": 3
}
],
"properties": {
"Node name for S&R": "DownloadAndLoadCogVideoModel"
},
"widgets_values": [
"THUDM/CogVideoX-2b",
"fp16"
]
},
{
"id": 32,
"type": "VHS_VideoCombine",
@ -323,7 +326,7 @@
],
"size": [
563.3333740234375,
686.2222493489583
310
],
"flags": {},
"order": 7,
@ -365,7 +368,7 @@
"widgets_values": {
"frame_rate": 8,
"loop_count": 0,
"filename_prefix": "AnimateDiff",
"filename_prefix": "CogVideo2B_long",
"format": "video/h264-mp4",
"pix_fmt": "yuv420p",
"crf": 19,
@ -376,7 +379,7 @@
"hidden": false,
"paused": false,
"params": {
"filename": "AnimateDiff_00002.mp4",
"filename": "AnimateDiff_00001.mp4",
"subfolder": "",
"type": "temp",
"format": "video/h264-mp4",
@ -464,10 +467,10 @@
"config": {},
"extra": {
"ds": {
"scale": 0.9090909090909091,
"scale": 0.8264462809917354,
"offset": [
49.8551278885073,
87.4070604693312
86.92928825501215,
77.5537144406024
]
}
},

681
examples/cogvideo_vid2vid_test_example_01.json → examples/cogvideo_2b_vid2vid_test_example_02.json
View File

@ -1,6 +1,6 @@
{
"last_node_id": 70,
"last_link_id": 181,
"last_node_id": 73,
"last_link_id": 191,
"nodes": [
{
"id": 20,
@ -24,8 +24,8 @@
54,
56
],
"shape": 3,
"slot_index": 0
"slot_index": 0,
"shape": 3
}
],
"properties": {
@ -36,64 +36,6 @@
"sd3"
]
},
{
"id": 55,
"type": "GetImageSizeAndCount",
"pos": [
1165,
119
],
"size": {
"0": 210,
"1": 86
},
"flags": {},
"order": 12,
"mode": 0,
"inputs": [
{
"name": "image",
"type": "IMAGE",
"link": 118,
"slot_index": 0
}
],
"outputs": [
{
"name": "image",
"type": "IMAGE",
"links": [
170
],
"shape": 3,
"slot_index": 0
},
{
"name": "720 width",
"type": "INT",
"links": null,
"shape": 3
},
{
"name": "480 height",
"type": "INT",
"links": null,
"shape": 3
},
{
"name": "26 count",
"type": "INT",
"links": [
121
],
"shape": 3,
"slot_index": 3
}
],
"properties": {
"Node name for S&R": "GetImageSizeAndCount"
}
},
{
"id": 31,
"type": "CogVideoTextEncode",
@ -122,8 +64,8 @@
"links": [
161
],
"shape": 3,
"slot_index": 0
"slot_index": 0,
"shape": 3
}
],
"properties": {
@ -133,39 +75,6 @@
""
]
},
{
"id": 1,
"type": "DownloadAndLoadCogVideoModel",
"pos": [
649,
182
],
"size": {
"0": 315,
"1": 58
},
"flags": {},
"order": 1,
"mode": 0,
"outputs": [
{
"name": "cogvideo_pipe",
"type": "COGVIDEOPIPE",
"links": [
83,
159
],
"shape": 3,
"slot_index": 0
}
],
"properties": {
"Node name for S&R": "DownloadAndLoadCogVideoModel"
},
"widgets_values": [
"bf16"
]
},
{
"id": 41,
"type": "ImageResizeKJ",
@ -215,8 +124,8 @@
"links": [
126
],
"shape": 3,
"slot_index": 0
"slot_index": 0,
"shape": 3
},
{
"name": "width",
@ -241,7 +150,8 @@
false,
2,
0,
0
0,
"disabled"
]
},
{
@ -279,244 +189,14 @@
"links": [
172
],
"shape": 3,
"slot_index": 0
"slot_index": 0,
"shape": 3
}
],
"properties": {
"Node name for S&R": "CogVideoImageEncode"
}
},
{
"id": 59,
"type": "GetImageRangeFromBatch",
"pos": [
1459,
413
],
"size": {
"0": 315,
"1": 102
},
"flags": {
"collapsed": true
},
"order": 14,
"mode": 0,
"inputs": [
{
"name": "images",
"type": "IMAGE",
"link": 136,
"slot_index": 0
},
{
"name": "masks",
"type": "MASK",
"link": null
},
{
"name": "num_frames",
"type": "INT",
"link": 133,
"widget": {
"name": "num_frames"
}
}
],
"outputs": [
{
"name": "IMAGE",
"type": "IMAGE",
"links": [
134
],
"shape": 3,
"slot_index": 0
},
{
"name": "MASK",
"type": "MASK",
"links": null,
"shape": 3
}
],
"properties": {
"Node name for S&R": "GetImageRangeFromBatch"
},
"widgets_values": [
0,
29
]
},
{
"id": 58,
"type": "ImageConcanate",
"pos": [
1451,
476
],
"size": {
"0": 315,
"1": 102
},
"flags": {},
"order": 15,
"mode": 0,
"inputs": [
{
"name": "image1",
"type": "IMAGE",
"link": 134
},
{
"name": "image2",
"type": "IMAGE",
"link": 170
}
],
"outputs": [
{
"name": "IMAGE",
"type": "IMAGE",
"links": [
132
],
"shape": 3,
"slot_index": 0
}
],
"properties": {
"Node name for S&R": "ImageConcanate"
},
"widgets_values": [
"right",
false
]
},
{
"id": 56,
"type": "SimpleMath+",
"pos": [
1454,
366
],
"size": {
"0": 315,
"1": 78
},
"flags": {
"collapsed": true
},
"order": 13,
"mode": 0,
"inputs": [
{
"name": "a",
"type": "INT,FLOAT",
"link": 121
},
{
"name": "b",
"type": "INT,FLOAT",
"link": 171
}
],
"outputs": [
{
"name": "INT",
"type": "INT",
"links": [
133
],
"shape": 3,
"slot_index": 0
},
{
"name": "FLOAT",
"type": "FLOAT",
"links": null,
"shape": 3
}
],
"properties": {
"Node name for S&R": "SimpleMath+"
},
"widgets_values": [
"a - b"
]
},
{
"id": 47,
"type": "VHS_VideoCombine",
"pos": [
1463,
-407
],
"size": [
1110,
711.3333333333333
],
"flags": {},
"order": 16,
"mode": 0,
"inputs": [
{
"name": "images",
"type": "IMAGE",
"link": 132
},
{
"name": "audio",
"type": "VHS_AUDIO",
"link": null
},
{
"name": "meta_batch",
"type": "VHS_BatchManager",
"link": null
},
{
"name": "vae",
"type": "VAE",
"link": null
}
],
"outputs": [
{
"name": "Filenames",
"type": "VHS_FILENAMES",
"links": null,
"shape": 3
}
],
"properties": {
"Node name for S&R": "VHS_VideoCombine"
},
"widgets_values": {
"frame_rate": 8,
"loop_count": 0,
"filename_prefix": "AnimateDiff",
"format": "video/nvenc_h264-mp4",
"pix_fmt": "yuv420p",
"bitrate": 10,
"megabit": true,
"save_metadata": true,
"pingpong": false,
"save_output": false,
"videopreview": {
"hidden": false,
"paused": false,
"params": {
"filename": "AnimateDiff_00007.mp4",
"subfolder": "",
"type": "temp",
"format": "video/nvenc_h264-mp4",
"frame_rate": 8
}
}
}
},
{
"id": 11,
"type": "CogVideoDecode",
@ -526,7 +206,7 @@
],
"size": {
"0": 210,
"1": 46
"1": 78
},
"flags": {},
"order": 11,
@ -550,13 +230,16 @@
"links": [
118
],
"shape": 3,
"slot_index": 0
"slot_index": 0,
"shape": 3
}
],
"properties": {
"Node name for S&R": "CogVideoDecode"
}
},
"widgets_values": [
false
]
},
{
"id": 30,
@ -586,8 +269,8 @@
"links": [
160
],
"shape": 3,
"slot_index": 0
"slot_index": 0,
"shape": 3
}
],
"properties": {
@ -625,10 +308,10 @@
"type": "IMAGE",
"links": [
129,
136
191
],
"shape": 3,
"slot_index": 0
"slot_index": 0,
"shape": 3
},
{
"name": "720 width",
@ -636,8 +319,8 @@
"links": [
165
],
"shape": 3,
"slot_index": 1
"slot_index": 1,
"shape": 3
},
{
"name": "480 height",
@ -645,19 +328,18 @@
"links": [
164
],
"shape": 3,
"slot_index": 2
"slot_index": 2,
"shape": 3
},
{
"name": "32 count",
"type": "INT",
"links": [
171,
178,
181
],
"shape": 3,
"slot_index": 3
"slot_index": 3,
"shape": 3
}
],
"properties": {
@ -705,8 +387,8 @@
"links": [
179
],
"shape": 3,
"slot_index": 0
"slot_index": 0,
"shape": 3
},
{
"name": "frame_count",
@ -784,29 +466,29 @@
"links": [
180
],
"shape": 3,
"slot_index": 0
"slot_index": 0,
"shape": 3
},
{
"name": "512 width",
"type": "INT",
"links": [],
"shape": 3,
"slot_index": 1
"slot_index": 1,
"shape": 3
},
{
"name": "256 height",
"type": "INT",
"links": [],
"shape": 3,
"slot_index": 2
"slot_index": 2,
"shape": 3
},
{
"name": "32 count",
"type": "INT",
"links": [],
"shape": 3,
"slot_index": 3
"slot_index": 3,
"shape": 3
}
],
"properties": {
@ -825,7 +507,7 @@
"1": 58
},
"flags": {},
"order": 2,
"order": 1,
"mode": 0,
"outputs": [
{
@ -854,10 +536,10 @@
1090,
290
],
"size": [
315,
370
],
"size": {
"0": 312.9841613769531,
"1": 342.8801574707031
},
"flags": {},
"order": 10,
"mode": 0,
@ -910,10 +592,10 @@
"name": "t_tile_length",
"type": "INT",
"link": 181,
"slot_index": 7,
"widget": {
"name": "t_tile_length"
},
"slot_index": 7
}
}
],
"outputs": [
@ -941,16 +623,221 @@
480,
720,
16,
8,
25,
6,
9,
13,
"fixed",
"DDIM",
32,
2,
0.8
"DDIM",
8,
0.85
]
},
{
"id": 1,
"type": "DownloadAndLoadCogVideoModel",
"pos": [
649,
182
],
"size": {
"0": 315,
"1": 82
},
"flags": {},
"order": 2,
"mode": 0,
"outputs": [
{
"name": "cogvideo_pipe",
"type": "COGVIDEOPIPE",
"links": [
83,
159
],
"slot_index": 0,
"shape": 3
}
],
"properties": {
"Node name for S&R": "DownloadAndLoadCogVideoModel"
},
"widgets_values": [
"THUDM/CogVideoX-2b",
"fp16"
]
},
{
"id": 58,
"type": "ImageConcanate",
"pos": [
1499,
433
],
"size": {
"0": 315,
"1": 102
},
"flags": {},
"order": 13,
"mode": 0,
"inputs": [
{
"name": "image1",
"type": "IMAGE",
"link": 191
},
{
"name": "image2",
"type": "IMAGE",
"link": 170
}
],
"outputs": [
{
"name": "IMAGE",
"type": "IMAGE",
"links": [
132
],
"slot_index": 0,
"shape": 3
}
],
"properties": {
"Node name for S&R": "ImageConcanate"
},
"widgets_values": [
"right",
false
]
},
{
"id": 55,
"type": "GetImageSizeAndCount",
"pos": [
1223,
122
],
"size": {
"0": 210,
"1": 86
},
"flags": {},
"order": 12,
"mode": 0,
"inputs": [
{
"name": "image",
"type": "IMAGE",
"link": 118,
"slot_index": 0
}
],
"outputs": [
{
"name": "image",
"type": "IMAGE",
"links": [
170
],
"slot_index": 0,
"shape": 3
},
{
"name": "720 width",
"type": "INT",
"links": null,
"shape": 3
},
{
"name": "480 height",
"type": "INT",
"links": null,
"shape": 3
},
{
"name": "32 count",
"type": "INT",
"links": [],
"slot_index": 3,
"shape": 3
}
],
"properties": {
"Node name for S&R": "GetImageSizeAndCount"
}
},
{
"id": 47,
"type": "VHS_VideoCombine",
"pos": [
1560,
-379
],
"size": [
1110,
711.3333333333333
],
"flags": {},
"order": 14,
"mode": 0,
"inputs": [
{
"name": "images",
"type": "IMAGE",
"link": 132
},
{
"name": "audio",
"type": "VHS_AUDIO",
"link": null
},
{
"name": "meta_batch",
"type": "VHS_BatchManager",
"link": null
},
{
"name": "vae",
"type": "VAE",
"link": null
}
],
"outputs": [
{
"name": "Filenames",
"type": "VHS_FILENAMES",
"links": null,
"shape": 3
}
],
"properties": {
"Node name for S&R": "VHS_VideoCombine"
},
"widgets_values": {
"frame_rate": 8,
"loop_count": 0,
"filename_prefix": "AnimateDiff",
"format": "video/nvenc_h264-mp4",
"pix_fmt": "yuv420p",
"bitrate": 10,
"megabit": true,
"save_metadata": true,
"pingpong": false,
"save_output": false,
"videopreview": {
"hidden": false,
"paused": false,
"params": {
"filename": "AnimateDiff_00001.mp4",
"subfolder": "",
"type": "temp",
"format": "video/nvenc_h264-mp4",
"frame_rate": 8
}
}
}
}
],
"links": [
@ -986,14 +873,6 @@
0,
"IMAGE"
],
[
121,
55,
3,
56,
0,
"INT,FLOAT"
],
[
126,
41,
@ -1018,30 +897,6 @@
0,
"IMAGE"
],
[
133,
56,
0,
59,
2,
"INT"
],
[
134,
59,
0,
58,
0,
"IMAGE"
],
[
136,
57,
0,
59,
0,
"IMAGE"
],
[
159,
1,
@ -1106,14 +961,6 @@
1,
"IMAGE"
],
[
171,
57,
3,
56,
1,
"INT,FLOAT"
],
[
172,
37,
@ -1161,16 +1008,24 @@
64,
7,
"INT"
],
[
191,
57,
0,
58,
0,
"IMAGE"
]
],
"groups": [],
"config": {},
"extra": {
"ds": {
"scale": 0.7513148009015777,
"scale": 0.620921323059155,
"offset": [
177.74090581831425,
461.56507330501444
298.59028824596885,
694.562497939138
]
}
},

413
examples/example_01.json
View File

@ -1,413 +0,0 @@
{
"last_node_id": 31,
"last_link_id": 57,
"nodes": [
{
"id": 28,
"type": "VHS_VideoCombine",
"pos": [
1432,
150
],
"size": [
667.752197265625,
310
],
"flags": {},
"order": 6,
"mode": 0,
"inputs": [
{
"name": "images",
"type": "IMAGE",
"link": 51,
"slot_index": 0
},
{
"name": "audio",
"type": "VHS_AUDIO",
"link": null
},
{
"name": "meta_batch",
"type": "VHS_BatchManager",
"link": null
},
{
"name": "vae",
"type": "VAE",
"link": null
}
],
"outputs": [
{
"name": "Filenames",
"type": "VHS_FILENAMES",
"links": null,
"shape": 3
}
],
"properties": {
"Node name for S&R": "VHS_VideoCombine"
},
"widgets_values": {
"frame_rate": 8,
"loop_count": 0,
"filename_prefix": "CogVideoX",
"format": "video/h264-mp4",
"pix_fmt": "yuv420p",
"crf": 19,
"save_metadata": true,
"pingpong": false,
"save_output": false,
"videopreview": {
"hidden": false,
"paused": false,
"params": {
"filename": "CogVideoX_00001.mp4",
"subfolder": "",
"type": "temp",
"format": "video/h264-mp4",
"frame_rate": 8
}
}
}
},
{
"id": 30,
"type": "CogVideoTextEncode",
"pos": [
500,
308
],
"size": {
"0": 474.8450012207031,
"1": 164.7423553466797
},
"flags": {},
"order": 2,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 54
}
],
"outputs": [
{
"name": "conditioning",
"type": "CONDITIONING",
"links": [
55
],
"shape": 3,
"slot_index": 0
}
],
"properties": {
"Node name for S&R": "CogVideoTextEncode"
},
"widgets_values": [
"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\nacoustic guitar, producing soft, melodic tunes. Nearby, a few other pandas gather, watching curiously and some clapping in rhythm. Sunlight filters\nthrough the tall bamboo, casting a gentle glow on the scene. The panda's face is expressive, showing concentration and joy as it plays. The\nbackground includes a small, flowing stream and vibrant green foliage, enhancing the peaceful and magical atmosphere of this unique musical\nperformance."
]
},
{
"id": 20,
"type": "CLIPLoader",
"pos": [
-59,
397
],
"size": {
"0": 451.30548095703125,
"1": 82
},
"flags": {},
"order": 0,
"mode": 0,
"outputs": [
{
"name": "CLIP",
"type": "CLIP",
"links": [
54,
56
],
"shape": 3,
"slot_index": 0
}
],
"properties": {
"Node name for S&R": "CLIPLoader"
},
"widgets_values": [
"t5\\google_t5-v1_1-xxl_encoderonly-fp8_e4m3fn.safetensors",
"sd3"
]
},
{
"id": 31,
"type": "CogVideoTextEncode",
"pos": [
503,
521
],
"size": {
"0": 463.01251220703125,
"1": 98.10446166992188
},
"flags": {},
"order": 3,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 56
}
],
"outputs": [
{
"name": "conditioning",
"type": "CONDITIONING",
"links": [
57
],
"shape": 3,
"slot_index": 0
}
],
"properties": {
"Node name for S&R": "CogVideoTextEncode"
},
"widgets_values": [
""
]
},
{
"id": 1,
"type": "DownloadAndLoadCogVideoModel",
"pos": [
649,
182
],
"size": {
"0": 315,
"1": 58
},
"flags": {},
"order": 1,
"mode": 0,
"outputs": [
{
"name": "cogvideo_pipe",
"type": "COGVIDEOPIPE",
"links": [
36
],
"shape": 3,
"slot_index": 0
}
],
"properties": {
"Node name for S&R": "DownloadAndLoadCogVideoModel"
},
"widgets_values": [
"bf16"
]
},
{
"id": 11,
"type": "CogVideoDecode",
"pos": [
1140,
783
],
"size": {
"0": 210,
"1": 46
},
"flags": {},
"order": 5,
"mode": 0,
"inputs": [
{
"name": "pipeline",
"type": "COGVIDEOPIPE",
"link": 37
},
{
"name": "samples",
"type": "LATENT",
"link": 38
}
],
"outputs": [
{
"name": "images",
"type": "IMAGE",
"links": [
51
],
"shape": 3,
"slot_index": 0
}
],
"properties": {
"Node name for S&R": "CogVideoDecode"
}
},
{
"id": 22,
"type": "CogVideoSampler",
"pos": [
1041,
342
],
"size": {
"0": 315,
"1": 382
},
"flags": {},
"order": 4,
"mode": 0,
"inputs": [
{
"name": "pipeline",
"type": "COGVIDEOPIPE",
"link": 36
},
{
"name": "positive",
"type": "CONDITIONING",
"link": 55,
"slot_index": 1
},
{
"name": "negative",
"type": "CONDITIONING",
"link": 57
},
{
"name": "samples",
"type": "LATENT",
"link": null
}
],
"outputs": [
{
"name": "cogvideo_pipe",
"type": "COGVIDEOPIPE",
"links": [
37
],
"shape": 3
},
{
"name": "samples",
"type": "LATENT",
"links": [
38
],
"shape": 3
}
],
"properties": {
"Node name for S&R": "CogVideoSampler"
},
"widgets_values": [
480,
720,
16,
8,
25,
6,
806286757407561,
"fixed",
"DDIM",
16,
2,
1
]
}
],
"links": [
[
36,
1,
0,
22,
0,
"COGVIDEOPIPE"
],
[
37,
22,
0,
11,
0,
"COGVIDEOPIPE"
],
[
38,
22,
1,
11,
1,
"LATENT"
],
[
51,
11,
0,
28,
0,
"IMAGE"
],
[
54,
20,
0,
30,
0,
"CLIP"
],
[
55,
30,
0,
22,
1,
"CONDITIONING"
],
[
56,
20,
0,
31,
0,
"CLIP"
],
[
57,
31,
0,
22,
2,
"CONDITIONING"
]
],
"groups": [],
"config": {},
"extra": {
"ds": {
"scale": 0.8264462809917356,
"offset": [
253.92700064075518,
186.82608107136124
]
}
},
"version": 0.4
}

1
nodes.py
View File

@ -188,7 +188,6 @@ class CogVideoImageEncode:
latents = vae.config.scaling_factor * latents
latents = latents.permute(0, 2, 1, 3, 4) # B, T_chunk, C, H, W
latents_list.append(latents)
vae.clear_fake_context_parallel_cache()
# Concatenate all the chunks along the temporal dimension
final_latents = torch.cat(latents_list, dim=1)