WAN 2.2 LONGER VIDEOS WORKFLOW!
Hey everyone! I've just released a quick new video update on GENERATING LONGER VIDEOS WITH WAN 2.2 !
Image to video, text to video AND EVEN video extend!😎 Now you can generate minutes long videos inside the same easy to use and modify workflow!
You can check out the video right here: https://youtu.be/ImJ32AlnM3A
I also created a 1-click installer for the Webui for my Patreon supporters for Local AND Runpod, that you can find right here: https://www.patreon.com/posts/135728905
Here the link to the manual install guide: https://pastebin.com/iuNPnrGm
Down below you will also find the WAN2_2-LONGER_VIDEOS_WORKFLOW.json for free for you to use!
As always, supporting me on Patreon allows me to keep creating helpful resources like this for the community. Thank you for your support - now go have some fun😉!
Comments (107)
Hi . I changed my computer and wanted to reinstall this as well as Z-Image, but after doing a clean install, opening your workflow I get the notification that nodes are missing and theres a list, if I click to install them, nothing happens and there's a red notification in the ComfyUI saying Comfy is outdated. If I try to update it using the manager, It says I dont have the required permission. I go to the config.ini and change the security value from strong to weak but it doesn't budge. Any recommendations? Same thing with Z-Image installer but different nodes missing. WAN Animate works, with some warnings. Is there a way you could future proof these installers so that they update and work even in couple of months in the future? Thanks.
In the readme it tells you that there's a specific update bat that allows you to update comfyui. which is neccesary to update the nodes due to security reasons. After you update comfyui with that bat file you can then install and update nodes as usual.
I tried chaining the gens. Each new chain gets darker and the video gets more purple. Is this normal?
How do you copy the nodes for further segments? I can eliminate the extras I don't need easily enough but when trying to copy and paste additional segments, as shown in the video, my nodes are scattered all over the place and need to be manually moved AND the lora nodes disappear completely.
Have you a fix for this problem? I want an I2V. i let turn the model 180 degrees around in the first sequence. in the 2end sequence she should turn back but her face is completly diffrent.
Can someone help me? I'm kinda new at this. I totally got this entire thing working after a while, and the biggest problem I have is having to manually enter the dimensions of the initial video in the video2video workflow. If I have like 6 or something flows in there, it gets tiresome, and there is the possibility that I forget one and it messes up the entire video chain. Please, there has to be a way that either it AUTO detects the initial video size and changes the rest to match, or at the very least, you enter it once and then it auto changes the rest regardless of how many chains you have. This is a really good workflow, but it seems this one little part could be automated, and leaves a lot to mess up if you forget a chain.
looking at how many people have issues and how many people actually get responses. Seems a bit unbalanced.
I am getting a "KSamplerAdvanced Given groups=1, weight of size [5120, 36, 1, 2, 2], expected input[1, 16, 21, 90, 160] to have 36 channels, but got 16 channels instead" Tried my best to solve this one, but even with the limited help of chatGPT, my knowledge isn't enough. Any fixaround ? --------------- Still having the issue on the portable ConfyUI, but seems to be working on a full ComfyUI install.
Make sure you're using the EXACT models in the initial workflow and do not change them to something similar. They are designed to work together; even if you have a better model, use the one in the initial flow. no compromise.
If only there was a way to add VACE to the workflow. More control over each clip as longer videos tend to drift
like some others I am getting the: UnetLoaderGGUF cannot reshape array of size 26511568 into shape (5120,5440) I have a single 5090. I am doing text to image but this happens with text to video as well. I have the same issue in both the wan 2.2 load and the wan 2.2 extended video. I am using your initial setup recreating your image with no changes. I have not gone back to doing a complete reinstall. Below is the terminal box text: C:\ai\ComfyUI_windows_portable>.\python_embeded\python.exe -s ComfyUI\main.py --windows-standalone-build [START] Security scan [DONE] Security scan ## ComfyUI-Manager: installing dependencies done. ** ComfyUI startup time: 2025-09-01 09:18:18.533 ** Platform: Windows ** Python version: 3.12.10 (tags/v3.12.10:0cc8128, Apr 8 2025, 12:21:36) [MSC v.1943 64 bit (AMD64)] ** Python executable: C:\ai\ComfyUI_windows_portable\python_embeded\python.exe ** ComfyUI Path: C:\ai\ComfyUI_windows_portable\ComfyUI ** ComfyUI Base Folder Path: C:\ai\ComfyUI_windows_portable\ComfyUI ** User directory: C:\ai\ComfyUI_windows_portable\ComfyUI\user ** ComfyUI-Manager config path: C:\ai\ComfyUI_windows_portable\ComfyUI\user\default\ComfyUI-Manager\config.ini ** Log path: C:\ai\ComfyUI_windows_portable\ComfyUI\user\comfyui.log Prestartup times for custom nodes: 0.0 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\rgthree-comfy 0.0 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Easy-Use 2.0 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Manager Checkpoint files will always be loaded safely. Total VRAM 32607 MB, total RAM 65295 MB pytorch version: 2.8.0+cu128 Set vram state to: NORMAL_VRAM Device: cuda:0 NVIDIA GeForce RTX 5090 : cudaMallocAsync Using pytorch attention Python version: 3.12.10 (tags/v3.12.10:0cc8128, Apr 8 2025, 12:21:36) [MSC v.1943 64 bit (AMD64)] ComfyUI version: 0.3.56 ComfyUI frontend version: 1.25.11 [Prompt Server] web root: C:\ai\ComfyUI_windows_portable\python_embeded\Lib\site-packages\comfyui_frontend_package\static [ComfyUI-Easy-Use] server: v1.3.3 Loaded [ComfyUI-Easy-Use] web root: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Easy-Use\web_version/v2 Loaded ComfyUI-GGUF: Allowing full torch compile Traceback (most recent call last): File "C:\ai\ComfyUI_windows_portable\ComfyUI\nodes.py", line 2129, in load_custom_node module_spec.loader.exec_module(module) File "", line 999, in exec_module File "", line 488, in _call_with_frames_removed File "C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-hunyuanvideowrapper\__init__.py", line 1, in from .nodes import NODE_CLASS_MAPPINGS as NODES_CLASS, NODE_DISPLAY_NAME_MAPPINGS as NODES_DISPLAY File "C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-hunyuanvideowrapper\nodes.py", line 10, in from .hyvideo.text_encoder import TextEncoder File "C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-hunyuanvideowrapper\hyvideo\text_encoder\__init__.py", line 9, in from .processing_llava import LlavaProcessor File "C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-hunyuanvideowrapper\hyvideo\text_encoder\processing_llava.py", line 23, in from transformers.processing_utils import ProcessingKwargs, ProcessorMixin, Unpack, _validate_images_text_input_order ImportError: cannot import name '_validate_images_text_input_order' from 'transformers.processing_utils' (C:\ai\ComfyUI_windows_portable\python_embeded\Lib\site-packages\transformers\processing_utils.py) Cannot import C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-hunyuanvideowrapper module for custom nodes: cannot import name '_validate_images_text_input_order' from 'transformers.processing_utils' (C:\ai\ComfyUI_windows_portable\python_embeded\Lib\site-packages\transformers\processing_utils.py) ### Loading: ComfyUI-Manager (V3.36) [ComfyUI-Manager] network_mode: public ### ComfyUI Version: v0.3.56-6-g9b151559 | Released on '2025-08-31' __ ______ ____ ____ ____ _ _ \ \ / / _ \ / ___| __ _ _ __ ___ ___| _ \ _____ __/ ___(_)_ __| | \ \ / /| |_) | | _ / _` | '_ ` _ \ / _ \ | | |/ _ \ \ / / | _| | '__| | \ V / | _ <| |_| | (_| | | | | | | __/ |_| | __/\ V /| |_| | | | | | \_/ |_| \_\\____|\__,_|_| |_| |_|\___|____/ \___| \_/ \____|_|_| |_| 🎮 VRGameDevGirl custom nodes loaded successfully! 🎞️ [ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/model-list.json [ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/alter-list.json [ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/github-stats.json [ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/extension-node-map.json [ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/custom-node-list.json ------------------------------------------ Comfyroll Studio v1.76 : 175 Nodes Loaded ------------------------------------------ ** For changes, please see patch notes at https://github.com/Suzie1/ComfyUI_Comfyroll_CustomNodes/blob/main/Patch_Notes.md ** For help, please see the wiki at https://github.com/Suzie1/ComfyUI_Comfyroll_CustomNodes/wiki ------------------------------------------ [C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux] | INFO -> Using ckpts path: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux\ckpts [C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux] | INFO -> Using symlinks: False [C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux] | INFO -> Using ort providers: ['CUDAExecutionProvider', 'DirectMLExecutionProvider', 'OpenVINOExecutionProvider', 'ROCMExecutionProvider', 'CPUExecutionProvider', 'CoreMLExecutionProvider'] C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux\node_wrappers\dwpose.py:26: UserWarning: DWPose: Onnxruntime not found or doesn't come with acceleration providers, switch to OpenCV with CPU device. DWPose might run very slowly warnings.warn("DWPose: Onnxruntime not found or doesn't come with acceleration providers, switch to OpenCV with CPU device. DWPose might run very slowly") # 😺dzNodes: LayerStyle -> Cannot import name 'guidedFilter' from 'cv2.ximgproc' A few nodes cannot works properly, while most nodes are not affected. Please REINSTALL package 'opencv-contrib-python'. For detail refer to https://github.com/chflame163/ComfyUI_LayerStyle/issues/5 [tinyterraNodes] Loaded Using pytorch attention (RES4LYF) Init (RES4LYF) Importing beta samplers. (RES4LYF) Importing legacy samplers. [rgthree-comfy] Loaded 48 exciting nodes. 🎉 WAS Node Suite: OpenCV Python FFMPEG support is enabled WAS Node Suite Warning: `ffmpeg_bin_path` is not set in `C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\was-node-suite-comfyui\was_suite_config.json` config file. Will attempt to use system ffmpeg binaries if available. WAS Node Suite: Finished. Loaded 220 nodes successfully. "Every artist was first an amateur." - Ralph Waldo Emerson Import times for custom nodes: 0.0 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\websocket_image_save.py 0.0 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-vrgamedevgirl 0.0 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-mxtoolkit 0.0 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-MagCache 0.0 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-GGUF 0.0 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\wlsh_nodes 0.0 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-custom-scripts 0.0 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_ultimatesdupscale 0.0 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\Derfuu_ComfyUI_ModdedNodes 0.0 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Frame-Interpolation 0.0 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-KJNodes 0.0 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\gguf 0.0 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux 0.0 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-VideoHelperSuite 0.1 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_tinyterranodes 0.1 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI_LayerStyle 0.1 seconds (IMPORT FAILED): C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-hunyuanvideowrapper 0.1 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\rgthree-comfy 0.1 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-advancedliveportrait 0.1 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Manager 0.2 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\RES4LYF 0.2 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-liveportraitkj 0.2 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI_essentials 0.4 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI_Comfyroll_CustomNodes 0.8 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\was-node-suite-comfyui 1.4 seconds: C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Easy-Use Context impl SQLiteImpl. Will assume non-transactional DDL. No target revision found. Starting server To see the GUI go to: http://127.0.0.1:8188 FETCH ComfyRegistry Data: 5/96 FETCH ComfyRegistry Data: 10/96 got prompt Using pytorch attention in VAE Using pytorch attention in VAE VAE load device: cuda:0, offload device: cpu, dtype: torch.bfloat16 FETCH ComfyRegistry Data: 15/96 FETCH ComfyRegistry Data: 20/96 gguf qtypes: Q5_K (168), F32 (73), Q6_K (1) Attempting to recreate sentencepiece tokenizer from GGUF file metadata... FETCH ComfyRegistry Data: 25/96 FETCH ComfyRegistry Data: 30/96 Created tokenizer with vocab size of 256384 Dequantizing token_embd.weight to prevent runtime OOM. CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cpu, dtype: torch.float16 gguf qtypes: F16 (694), Q8_0 (400), F32 (1) model weight dtype torch.float16, manual cast: None model_type FLOW Requested to load WanTEModel FETCH ComfyRegistry Data: 35/96 loaded completely 29323.8 5039.953125 True Requested to load WAN21 FETCH ComfyRegistry Data: 40/96 FETCH ComfyRegistry Data: 45/96 loaded partially 14529.994980926513 14529.990356445312 0 Attempting to release mmap (174) FETCH ComfyRegistry Data: 50/96 0%| | 0/2 [00:00
adding last part: ETCH ComfyRegistry Data: 60/96 FETCH ComfyRegistry Data: 65/96 FETCH ComfyRegistry Data: 70/96 FETCH ComfyRegistry Data: 75/96 FETCH ComfyRegistry Data: 80/96 FETCH ComfyRegistry Data: 85/96 FETCH ComfyRegistry Data: 90/96 FETCH ComfyRegistry Data: 95/96 FETCH ComfyRegistry Data [DONE] [ComfyUI-Manager] default cache updated: https://api.comfy.org/nodes FETCH DATA from: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/custom-node-list.json [DONE] [ComfyUI-Manager] All startup tasks have been completed. 100%|████████████████████████████████████████████████████████████████████████████████████| 2/2 [01:24<00:00, 42.38s/it] !!! Exception during processing !!! cannot reshape array of size 26511568 into shape (5120,5440) Traceback (most recent call last): File "C:\ai\ComfyUI_windows_portable\ComfyUI\execution.py", line 496, in execute output_data, output_ui, has_subgraph, has_pending_tasks = await get_output_data(prompt_id, unique_id, obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, hidden_inputs=hidden_inputs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\ai\ComfyUI_windows_portable\ComfyUI\execution.py", line 315, in get_output_data return_values = await _async_map_node_over_list(prompt_id, unique_id, obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, hidden_inputs=hidden_inputs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\ai\ComfyUI_windows_portable\ComfyUI\execution.py", line 289, in _async_map_node_over_list await process_inputs(input_dict, i) File "C:\ai\ComfyUI_windows_portable\ComfyUI\execution.py", line 277, in process_inputs result = f(**inputs) ^^^^^^^^^^^ File "C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-GGUF\nodes.py", line 152, in load_unet sd = gguf_sd_loader(unet_path) ^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-GGUF\loader.py", line 55, in gguf_sd_loader reader = gguf.GGUFReader(path) ^^^^^^^^^^^^^^^^^^^^^ File "C:\ai\ComfyUI_windows_portable\python_embeded\Lib\site-packages\gguf\gguf_reader.py", line 182, in __init__ self._build_tensors(offs, tensors_fields) File "C:\ai\ComfyUI_windows_portable\python_embeded\Lib\site-packages\gguf\gguf_reader.py", line 364, in _build_tensors data = self._get(data_offs, item_type, item_count).reshape(np_dims), ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ ValueError: cannot reshape array of size 26511568 into shape (5120,5440) Prompt executed in 117.21 seconds
question, can the two samplers be on random or do they have to be the same?
Random is fine
I need to apply wan video NAG so I can make the characters stop talking. Where do I have to put it on the workflow?
it works... kinda... it keeps the overall composition of the initial image but shifts it into a completely different style and ignores the character designs. So it basically just uses the initial image for placing objects and deciding on the pose without actually just turning the image into a video how I thought this would work... edit: nevermind... I accidentally put the T2V in the highnoise one... so of course it didn't work :-P
You can add nodes to correct this by running them through nano banana
Uhm.... yeah this is simply the "first frame last frame" technique. Problem is character drift. When the characters face is not well visible in the last frame or the eyes are closed... well... the ai will imagine a new face and you will have an ugly drift. I was hoping that worklflow would be more advanced but yeah.. its pretty basic
when i copy paste the nodes and connect them correctly to extend my generation (as you show on your video) the generation starts from the very beginning. it shouldn't touch the first video nodes if i didn't change any parameters right? so what is the issue here i don't know. i tried the bypass earlier generations but this time the last frame wasn't pulled..
mine goes very blurry for some reason
I'm having a weird bug. Every video in the chain become more and more red. Any ideas what could cause that ? Thanks
Is there a way to clear VHS_VideoCombine previews from last generation when I restart Comfyui?
what types of generation times is everyone getting?
Is there a node that let's you resize to a different aspect ratio specifically for 9:16 for image to video
Dear Aiteepreneur, would this be possible with image to image to image to image?
amazing!
Thank you! awesome. I will probably use my RTX crypto-miner rig to render faster videos.
hello i install but when i tried it doesnt do the job. im trying the long videos and the combo. when i make 1 click install do i need to do everytime that for each version? or when i have it installed all i need its to put the json acording? thank you
WanVideoNAG is missing, now i cant use any workflow after installing this. Everything was workign fine before this. please help. something wrong with the KJNodes for ComfyUI
is there a way to restart a generation from a previous step without rerunning the whole workflow. say if for example the previous video steps look good but the newest didn't work out
I’ve been following your work because you stand out in the community for easing the hassle of onboarding things, which is greatly appreciated. Full disclosure, My appreciation despite not being a paying member is fostered by this source I discovered a couple years ago that, very soon after you drop your videos, freely hosts all the content that you say is available here to your patrons (and a seriously huge selection of other creators on this platform as well), which certainly must hinder your income potential. From what I can tell, there's 807 people who've "favorited" you or whatever on that source. I figured that you might be interested perhaps to be informed of this source so that you could DMCA your content and force people like myself into a situation of having to pay you for the things we appreciate you for...in exchange for a small token of appreciation, of course. Please feel free to reach out if you care to discuss this further! pewpypewsie@gmail.com
I'm sure he knows about the party bud
I see in the 'log' this line: Requested to load WAN21 . Shouldn't that be WAN22? Anyone else seeing that?
should be loaded wan2.1 is correct
It's great. I wonder if it could be customized to help make AMV s.
Don't know if I am doing something wrong, but I can never get the second prompt to begin from the last image output. I am using the WAN2.2 Q4_K_S variant?, also my videos tend to cycle forward and then backwards to their start position. Has anyone else run into these issues?
Same here. The GPU keeps spinning up but the the second video never comes out. I'm on Q4 as well, spinning 3080 RTX - Text to Video.
What are some good prompt generators/guides for Wan 2.2 text-to-video videos?
Maybe try this: https://alidocs.dingtalk.com/i/nodes/EpGBa2Lm8aZxe5myC99MelA2WgN7R35y
on runpod setup i get a weird bug that after a few runs it get stuck on images batch multi and memory gets stuck at 57.74GiB / 57.74GiB. need to stop my pod en reconnect after as a reboot. however when trying to run it again i get the same error at the final part.
I have the same issue, I have spent 10$ on runpod without getting anything done because of this
Text to video just outputs a bunch of gray fuzz?
Hey K, can we please get a 3 image input like your 1st 1 for pulid but for nunchaku pulid?
anyone else having videos get increasingly blurry with I2V?
yeah, once it hits the 4th node, it stars to blur and slows the output frames and then the last set it just a blurry mess. I get about 8 seconds of useable footage.
Not sure why I can't get this to work, the lightning loras have no effect at all, took ten minutes to generate a poor quality 6 second output
It is clear that the tools you develop for ComfyUI are well thought out and are of a superb caliber, demonstrating your masterful skillset. Thank you for all your efforts and continuous success! Before running the WAN2_2-ULTRA-MODELS-NODES_INSTALL.bat file, I backed up my root ComfyUI directory to ensure nothing would break. I had previously spent a lot of time compiling Torch for CUDA 12.8, which worked with my RTX 3090 24GB, along with compatible versions of xformers and Sage Attention. After running the .bat file and choosing option 2 for the 12-24GB RAM GPU, I attempted to run ComfyUI but found that it had updated to CUDA 12.9, which Torch was not compiled for and could no longer function without recompiling Torch. Are there steps in the .bat file that can be skipped or modified, or is there another way for me to retain PyTorch version 2.7.1+cu12.8, xformers version 0.0.32+baf5eb1f.d20250710, and Python version 3.12.10 while still downloading/updating the required models for this workflow to work as intended? Below is a log of my output from starting up ComfyUI prior to running the .bat file for reference: ## ComfyUI-Manager: installing dependencies done. [2025-08-16 17:55:42.167] ** ComfyUI startup time: 2025-08-16 17:55:42.167 [2025-08-16 17:55:42.167] ** Platform: Windows [2025-08-16 17:55:42.167] ** Python version: 3.12.10 (tags/v3.12.10:0cc8128, Apr 8 2025, 12:21:36) [MSC v.1943 64 bit (AMD64)] [2025-08-16 17:55:42.168] ** Python executable: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\python_embeded\python.exe [2025-08-16 17:55:42.168] ** ComfyUI Path: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI [2025-08-16 17:55:42.168] ** ComfyUI Base Folder Path: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI [2025-08-16 17:55:42.168] ** User directory: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\user [2025-08-16 17:55:42.168] ** ComfyUI-Manager config path: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\user\default\ComfyUI-Manager\config.ini [2025-08-16 17:55:42.169] ** Log path: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\user\comfyui.log Prestartup times for custom nodes: [2025-08-16 17:55:43.169] 0.0 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\rgthree-comfy [2025-08-16 17:55:43.175] 0.0 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Easy-Use [2025-08-16 17:55:43.176] 2.5 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Manager [2025-08-16 17:55:43.176] [2025-08-16 17:55:56.301] Checkpoint files will always be loaded safely. [2025-08-16 17:55:56.389] Total VRAM 24575 MB, total RAM 32577 MB [2025-08-16 17:55:56.390] pytorch version: 2.7.1+cu128 [2025-08-16 17:55:58.676] xformers version: 0.0.32+baf5eb1f.d20250710 [2025-08-16 17:55:58.676] Set vram state to: NORMAL_VRAM [2025-08-16 17:55:58.676] Device: cuda:0 NVIDIA GeForce RTX 3090 : cudaMallocAsync [2025-08-16 17:55:58.989] Using xformers attention [2025-08-16 17:56:01.441] Python version: 3.12.10 (tags/v3.12.10:0cc8128, Apr 8 2025, 12:21:36) [MSC v.1943 64 bit (AMD64)] [2025-08-16 17:56:01.441] ComfyUI version: 0.3.43 [2025-08-16 17:56:01.566] ComfyUI frontend version: 1.23.4 [2025-08-16 17:56:01.569] [Prompt Server] web root: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\python_embeded\Lib\site-packages\comfyui_frontend_package\static [2025-08-16 17:56:03.044] [Crystools ?[0;32mINFO?[0m] Crystools version: 1.25.2 [2025-08-16 17:56:03.082] [Crystools ?[0;32mINFO?[0m] Platform release: 11 [2025-08-16 17:56:03.082] [Crystools ?[0;32mINFO?[0m] JETSON: Not detected. [2025-08-16 17:56:03.083] [Crystools ?[0;32mINFO?[0m] CPU: 12th Gen Intel(R) Core(TM) i7-12700K - Arch: AMD64 - OS: Windows 11 [2025-08-16 17:56:03.093] [Crystools ?[0;32mINFO?[0m] pynvml (NVIDIA) initialized. [2025-08-16 17:56:03.093] [Crystools ?[0;32mINFO?[0m] GPU/s: [2025-08-16 17:56:03.108] [Crystools ?[0;32mINFO?[0m] 0) NVIDIA GeForce RTX 3090 [2025-08-16 17:56:03.124] [Crystools ?[0;32mINFO?[0m] 1) NVIDIA GeForce RTX 3070 [2025-08-16 17:56:03.124] [Crystools ?[0;32mINFO?[0m] NVIDIA Driver: 577.00 [2025-08-16 17:56:05.736] ?[34m[ComfyUI-Easy-Use] server: ?[0mv1.3.2 ?[92mLoaded?[0m [2025-08-16 17:56:05.736] ?[34m[ComfyUI-Easy-Use] web root: ?[0mC:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Easy-Use\web_version/v2 ?[92mLoaded?[0m [2025-08-16 17:56:05.756] ComfyUI-GGUF: Partial torch compile only, consider updating pytorch [2025-08-16 17:56:05.760] ### Loading: ComfyUI-Impact-Pack (V8.21.1) [2025-08-16 17:56:06.077] [Impact Pack] Wildcards loading done. [2025-08-16 17:56:06.388] ### Loading: ComfyUI-Manager (V3.35) [2025-08-16 17:56:06.389] [ComfyUI-Manager] network_mode: public [2025-08-16 17:56:06.523] ### ComfyUI Revision: 150 [e18f53cc] *DETACHED | Released on '2025-06-27' [2025-08-16 17:56:06.554] MultiGPU: Checking for module at C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-LTXVideo [2025-08-16 17:56:06.554] MultiGPU: Module ComfyUI-LTXVideo not found - skipping [2025-08-16 17:56:06.555] MultiGPU: Checking for module at C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-ltxvideo [2025-08-16 17:56:06.555] MultiGPU: Module comfyui-ltxvideo not found - skipping [2025-08-16 17:56:06.555] MultiGPU: Checking for module at C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Florence2 [2025-08-16 17:56:06.555] MultiGPU: Module ComfyUI-Florence2 not found - skipping [2025-08-16 17:56:06.555] MultiGPU: Checking for module at C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-florence2 [2025-08-16 17:56:06.556] MultiGPU: Module comfyui-florence2 not found - skipping [2025-08-16 17:56:06.556] MultiGPU: Checking for module at C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI_bitsandbytes_NF4 [2025-08-16 17:56:06.556] MultiGPU: Module ComfyUI_bitsandbytes_NF4 not found - skipping [2025-08-16 17:56:06.556] MultiGPU: Checking for module at C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_bitsandbytes_nf4 [2025-08-16 17:56:06.556] MultiGPU: Module comfyui_bitsandbytes_nf4 not found - skipping [2025-08-16 17:56:06.556] MultiGPU: Checking for module at C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\x-flux-comfyui [2025-08-16 17:56:06.556] MultiGPU: Module x-flux-comfyui not found - skipping [2025-08-16 17:56:06.556] MultiGPU: Checking for module at C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\x-flux-comfyui [2025-08-16 17:56:06.556] MultiGPU: Module x-flux-comfyui not found - skipping [2025-08-16 17:56:06.556] MultiGPU: Checking for module at C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-MMAudio [2025-08-16 17:56:06.557] MultiGPU: Module ComfyUI-MMAudio not found - skipping [2025-08-16 17:56:06.557] MultiGPU: Checking for module at C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-mmaudio [2025-08-16 17:56:06.557] MultiGPU: Module comfyui-mmaudio not found - skipping [2025-08-16 17:56:06.557] MultiGPU: Checking for module at C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-GGUF [2025-08-16 17:56:06.557] MultiGPU: Found ComfyUI-GGUF, creating compatible MultiGPU nodes [2025-08-16 17:56:06.557] MultiGPU: Checking for module at C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\PuLID_ComfyUI [2025-08-16 17:56:06.557] MultiGPU: Module PuLID_ComfyUI not found - skipping [2025-08-16 17:56:06.557] MultiGPU: Checking for module at C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\pulid_comfyui [2025-08-16 17:56:06.557] MultiGPU: Module pulid_comfyui not found - skipping [2025-08-16 17:56:06.557] MultiGPU: Checking for module at C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-HunyuanVideoWrapper [2025-08-16 17:56:06.557] MultiGPU: Module ComfyUI-HunyuanVideoWrapper not found - skipping [2025-08-16 17:56:06.557] MultiGPU: Checking for module at C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-hunyuanvideowrapper [2025-08-16 17:56:06.558] MultiGPU: Module comfyui-hunyuanvideowrapper not found - skipping [2025-08-16 17:56:06.558] MultiGPU: Checking for module at C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-WanVideoWrapper [2025-08-16 17:56:06.558] MultiGPU: Found ComfyUI-WanVideoWrapper, creating compatible MultiGPU nodes [2025-08-16 17:56:06.558] MultiGPU: Registration complete. Final mappings: DeviceSelectorMultiGPU, HunyuanVideoEmbeddingsAdapter, MergeFluxLoRAsQuantizeAndLoaddMultiGPU, UNETLoaderMultiGPU, VAELoaderMultiGPU, CLIPLoaderMultiGPU, DualCLIPLoaderMultiGPU, TripleCLIPLoaderMultiGPU, QuadrupleCLIPLoaderMultiGPU, CLIPVisionLoaderMultiGPU, CheckpointLoaderSimpleMultiGPU, ControlNetLoaderMultiGPU, UnetLoaderGGUFMultiGPU, UnetLoaderGGUFDisTorchMultiGPU, UnetLoaderGGUFAdvancedMultiGPU, UnetLoaderGGUFAdvancedDisTorchMultiGPU, CLIPLoaderGGUFMultiGPU, CLIPLoaderGGUFDisTorchMultiGPU, DualCLIPLoaderGGUFMultiGPU, DualCLIPLoaderGGUFDisTorchMultiGPU, TripleCLIPLoaderGGUFMultiGPU, TripleCLIPLoaderGGUFDisTorchMultiGPU, QuadrupleCLIPLoaderGGUFMultiGPU, QuadrupleCLIPLoaderGGUFDisTorchMultiGPU, WanVideoModelLoaderMultiGPU, WanVideoVAELoaderMultiGPU, LoadWanVideoT5TextEncoderMultiGPU [2025-08-16 17:56:06.570] ======================================== ComfyUI-nunchaku Initialization ======================================== [2025-08-16 17:56:06.572] Nunchaku version: 0.3.1.dev20250609 [2025-08-16 17:56:06.573] ComfyUI-nunchaku version: 0.3.3 [2025-08-16 17:56:06.881] [ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/model-list.json [2025-08-16 17:56:06.882] [ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/alter-list.json [2025-08-16 17:56:07.007] [ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/github-stats.json [2025-08-16 17:56:07.062] [ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/extension-node-map.json [2025-08-16 17:56:07.115] [ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/custom-node-list.json [2025-08-16 17:56:08.270] C:\ComfyUI KONTEXT\ComfyUI_windows_portable\python_embeded\Lib\site-packages\timm\models\layers\__init__.py:48: FutureWarning: Importing from timm.models.layers is deprecated, please import via timm.layers warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning) [2025-08-16 17:56:08.275] Please 'pip install apex' [2025-08-16 17:56:08.282] ================================================================================================================= [2025-08-16 17:56:08.873] ?[0;33m[ReActor]?[0m - ?[38;5;173mSTATUS?[0m - ?[0;32mRunning v0.6.1-b2 in ComfyUI?[0m [2025-08-16 17:56:08.991] Torch version: 2.7.1+cu128 [2025-08-16 17:56:10.284] Traceback (most recent call last): File "C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\nodes.py", line 2124, in load_custom_node module_spec.loader.exec_module(module) File "", line 999, in exec_module File "", line 488, in _call_with_frames_removed File "C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Upscaler-Tensorrt\__init__.py", line 6, in from .trt_utilities import Engine File "C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Upscaler-Tensorrt\trt_utilities.py", line 32, in import tensorrt as trt ModuleNotFoundError: No module named 'tensorrt' [2025-08-16 17:56:10.284] Cannot import C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Upscaler-Tensorrt module for custom nodes: No module named 'tensorrt' [2025-08-16 17:56:10.538] __ ______ ____ ____ ____ _ _ \ \ / / _ \ / ___| __ _ _ __ ___ ___| _ \ _____ __/ ___(_)_ __| | \ \ / /| |_) | | _ / _` | '_ ` _ \ / _ \ | | |/ _ \ \ / / | _| | '__| | \ V / | _ <| |_| | (_| | | | | | | __/ |_| | __/\ V /| |_| | | | | | \_/ |_| \_\\____|\__,_|_| |_| |_|\___|____/ \___| \_/ \____|_|_| |_| 🎮 VRGameDevGirl custom nodes loaded successfully! 🎞️ [2025-08-16 17:56:10.539] [2025-08-16 17:56:11.239] FETCH ComfyRegistry Data: 5/94 [2025-08-16 17:56:11.502] ?[36;20m[C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux] | INFO -> Using ckpts path: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux\ckpts?[0m [2025-08-16 17:56:11.502] ?[36;20m[C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux] | INFO -> Using symlinks: False?[0m [2025-08-16 17:56:11.502] ?[36;20m[C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux] | INFO -> Using ort providers: ['CUDAExecutionProvider', 'DirectMLExecutionProvider', 'OpenVINOExecutionProvider', 'ROCMExecutionProvider', 'CPUExecutionProvider', 'CoreMLExecutionProvider']?[0m [2025-08-16 17:56:11.555] ?[34mIb Custom Nodes: ?[92mLoaded?[0m [2025-08-16 17:56:11.566] # 😺dzNodes: LayerStyle -> ?[1;33mCannot import name 'guidedFilter' from 'cv2.ximgproc' A few nodes cannot works properly, while most nodes are not affected. Please REINSTALL package 'opencv-contrib-python'. For detail refer to ?[4mhttps://github.com/chflame163/ComfyUI_LayerStyle/issues/5?[0m?[m [2025-08-16 17:56:11.951] [2025-08-16 17:56:11.952] ?[92m[rgthree-comfy] Loaded 47 epic nodes. 🎉?[0m [2025-08-16 17:56:11.953] [2025-08-16 17:56:11.958] Import times for custom nodes: [2025-08-16 17:56:11.958] 0.0 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\websocket_image_save.py [2025-08-16 17:56:11.958] 0.0 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-vrgamedevgirl [2025-08-16 17:56:11.958] 0.0 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\wlsh_nodes [2025-08-16 17:56:11.958] 0.0 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI_Ib_CustomNodes [2025-08-16 17:56:11.958] 0.0 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-GGUF [2025-08-16 17:56:11.958] 0.0 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-MagCache [2025-08-16 17:56:11.958] 0.0 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI_essentials [2025-08-16 17:56:11.958] 0.0 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-multigpu [2025-08-16 17:56:11.958] 0.0 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-custom-scripts [2025-08-16 17:56:11.958] 0.0 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI_UltimateSDUpscale [2025-08-16 17:56:11.959] 0.0 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-NAG [2025-08-16 17:56:11.959] 0.0 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Frame-Interpolation [2025-08-16 17:56:11.959] 0.0 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\Derfuu_ComfyUI_ModdedNodes [2025-08-16 17:56:11.959] 0.0 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\gguf [2025-08-16 17:56:11.959] 0.0 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-KJNodes [2025-08-16 17:56:11.959] 0.0 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux [2025-08-16 17:56:11.959] 0.1 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI_LayerStyle [2025-08-16 17:56:11.959] 0.2 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Manager [2025-08-16 17:56:11.959] 0.3 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-VideoHelperSuite [2025-08-16 17:56:11.959] 0.3 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\rgthree-comfy [2025-08-16 17:56:11.959] 0.3 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-logicutils [2025-08-16 17:56:11.959] 0.3 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Impact-Pack [2025-08-16 17:56:11.959] 0.3 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-reactor [2025-08-16 17:56:11.960] 0.5 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Crystools [2025-08-16 17:56:11.960] 0.6 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-ollama [2025-08-16 17:56:11.960] 1.0 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-WanVideoWrapper [2025-08-16 17:56:11.960] 1.1 seconds (IMPORT FAILED): C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Upscaler-Tensorrt [2025-08-16 17:56:11.960] 1.7 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-nunchaku [2025-08-16 17:56:11.960] 2.6 seconds: C:\ComfyUI KONTEXT\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Easy-Use [2025-08-16 17:56:11.960] [2025-08-16 17:56:12.241] Context impl SQLiteImpl. [2025-08-16 17:56:12.241] Will assume non-transactional DDL. [2025-08-16 17:56:12.242] No target revision found. [2025-08-16 17:56:12.263] Starting server [2025-08-16 17:56:12.263] To see the GUI go to: http://127.0.0.1:8188 [2025-08-16 17:56:15.789] FETCH ComfyRegistry Data: 10/94 [2025-08-16 17:56:19.764] FETCH ComfyRegistry Data: 15/94 [2025-08-16 17:56:23.746] FETCH ComfyRegistry Data: 20/94 [2025-08-16 17:56:28.034] FETCH ComfyRegistry Data: 25/94 [2025-08-16 17:56:32.007] FETCH ComfyRegistry Data: 30/94 [2025-08-16 17:56:35.943] FETCH ComfyRegistry Data: 35/94 [2025-08-16 17:56:39.955] FETCH ComfyRegistry Data: 40/94 [2025-08-16 17:56:43.885] FETCH ComfyRegistry Data: 45/94 [2025-08-16 17:56:47.886] FETCH ComfyRegistry Data: 50/94 [2025-08-16 17:56:51.815] FETCH ComfyRegistry Data: 55/94 [2025-08-16 17:56:55.740] FETCH ComfyRegistry Data: 60/94 [2025-08-16 17:56:59.805] FETCH ComfyRegistry Data: 65/94 [2025-08-16 17:57:03.821] FETCH ComfyRegistry Data: 70/94 [2025-08-16 17:57:07.909] FETCH ComfyRegistry Data: 75/94 [2025-08-16 17:57:11.862] FETCH ComfyRegistry Data: 80/94 [2025-08-16 17:57:15.794] FETCH ComfyRegistry Data: 85/94 [2025-08-16 17:57:20.167] FETCH ComfyRegistry Data: 90/94 [2025-08-16 17:57:24.687] FETCH ComfyRegistry Data [DONE] [2025-08-16 17:57:24.864] [ComfyUI-Manager] default cache updated: https://api.comfy.org/nodes [2025-08-16 17:57:24.896] FETCH DATA from: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/custom-node-list.json [DONE] [2025-08-16 17:57:25.264] [ComfyUI-Manager] All startup tasks have been completed. [2025-08-16 18:40:59.410] HTTP Request: GET http://127.0.0.1:11434/api/tags "HTTP/1.1 200 OK" [2025-08-16 18:40:59.465] Exception in callback _ProactorBasePipeTransport._call_connection_lost(None) handle:
Working perfectly out of the box :)
Excellent workflow! Can this workflow be explored further using a First Image + Last Image chain? The idea having more control over the scene by fusing key images together you want every few seconds. This will allow for precise choreography for things like fight scenes and intense action sequences.
Works great. The only thing I miss is that I can set everything once in the first video segment and don't have to set everything 5 times. Like video width/height, lenght, framerate, steps, etc. But anyway, good idea and can save a lot of time.
Yes, I wish the common controls could be synced to update across the next parts when set on the first; it's a hassle to change these every time. Otherwise, it's great.
It's just not working i have been trying all day long the Videos are just muddy I even tried your json straight , just lowered the dimensions to make it not run out of RAM. https://imgur.com/a/sydSPaA
same here... workflow produces noisy mess. previous workflow worked fine.
I see that you use the GGUF versions of Wan in your workflows. Do you find them to have better quality than the fp8 scaled versions?
First I Just want a say thank you for this awesome workflow ! Do you think it would be possible to have one like that but for the "Frame To Frame" ?
That would be cool, but take up so much space, lol
I have a 4080 16GB even I run the 4Q version for quicker speed etc. Takes about 31 minutes to generate a 16 second Text to Video clip at vanilla workflow settings. My brother has a 5060 8GB. Do any of you even believe 4Q would for him at all? I know he can just download it and give it a try. But I suspect he can't? Any 8GB GPU owners find a way to get this to work for sure? And if so did you have to tweak anything or even get like a 3Q or something?
Oof 8g will work but it will take...hours...Just rend a gpu on runpod instead
This is awesome and is working pretty well. Thank you.
Hey Aitrepreneur, thank you again for the awesome workflow my guy :) Do you still recommend only using Lora on the Low Noise Lora section, unless they're made specifically for the Wan2.2 model, which in that case can be used in both the High & Low noise Lora addons? Again, appreciate your diligent efforts!
this is only for the initial text to video workflow, for the rest it all uses the image to video
Amazing, works great!! Any way to start part way through? For example, say I'm unhappy with the results of my 3rd prompt but the first 2 are fine. Can I have it start at 3 leaving the 1 and 2 videos intact?
I suppose a work around would be to take the video output of 1+2 as the video input to the 'video base to video' section and start re-prompting from there.
Using the 4090 gpu on runpod it doesn't complete the action and loses connection. GPT says because there is not enough system ram. I tried the 5090 but it didn't work. Do you know which others could work?
you can use any other gpu like A40 for example
Nice work! I'm going to modify this for kijai's wrapper, I like the control his nodes give me. I see you're trying to keep things simple for people so they don't have to troubleshoot, but I'll say kijai's CFG scheduler that lets you add a higher CFG to the first step of each sampler is pretty fantastic for improving prompt adherence with how wacky lightx2v makes things. It adds a little time but you get much better results. It's in his i2v example workflow, and it should work with native as well.
could you share it with us or point to the right direction? im getting still videos on the first frame and sometimes completly other things than the input image
That awkward moment when you can't seem to get Torch to initialize, and so you check your torch version, only to find it's not installed.... How--...... Hey Aitrepeneur, could you drop the cmd to install the latest version for my 5070ti? I think I found it on google, but want to make sure. Also, does the cmd have to be run in a certain folder? If so, where?
It shouldn't be a problem, it's a portable install, everything is already inside the install folder. What exactly is the error?
i have 12gb vram and i'm getting a torch out of memory error , i'm using the Q4_K_S ... i'm wondering what it could be ?!
Wan2.2 is pretty intensive so you could definitely be getting out of memory due to that, even with a quantized model.
Encountered an issue with Fast Groups Bypasser node not showing any on/off toggle, not just for this workflow but for others. If you're having a similar issue try going into the Custom Node Manager, search for "rgthree" and switch to a previous version. 1.0.2507072340 seems to be working for me. Great WF btw! Love your work.
you can just use the update comfyui bat file inside the update folder to update comfyui correctly and use the update all to update all the nodes, then restart and redrag the workflow inside again
This is great! Please consider using sub graph nodes so there is only one input and one output node. Makes duplicating a lot easier
once again, it's cleaner but more difficult to help people troubleshoot issues, I think of everyone here
Nice WF, but you shoud really get comfortable with the set- and get-node. Makes things much easier.
cleaner maybe, it's just more annoying when you need to troubleshoot issues
Thank you! This solves some of the problems Framepack had solved for me, but the quality on this is just better! I appreciate your effort, it saves me a lot of time with the workflows.
okay I'm struggling with this one. I tested the first Image to Video workflow THREE times, Wan doesn't care about my input image at all and create something else on its own, completely different :/ First test I forgot to change the resolution, second test I used 720x1280, 3rd test I tried the same resolution as the input 832x1216. No luck, it doesn't work at all
you are probably using the text to video model instead of the image to video, many makes the same mistake
never did find your models, a link to where to down load the models in these would be AWESOME.. please and thank you..
Just use the installer, everything will be installed automatically, or you have the link and name of the models in the manual install link as well
Works Perfectly! Thank you!
i´m not sure what the problem is but i can´t run any of your 1 click installer. every cmd window has "oco" in the list. for example : Which QWEN IMAGE quantization? 1) Q4_K_S ÔÇô GPUs < 12 GB VRAM RECOMMENDED 2) Q5_K_S ÔÇô GPUs 12-24 GB 3) Q8_0 ÔÇô Best quality, GPUs 24 GB+ Enter 1, 2 or 3: when i try it then : Git not found ÔÇô downloading silent installer... % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 100 9 100 9 0 0 41 0 --:--:-- --:--:-- --:--:-- 41
you don't have git for windows installed
Heyy Im using runpod but getting this error CLIPTextEncode CUDA error: no kernel image is available for execution on the device CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect. For debugging consider passing CUDA_LAUNCH_BLOCKING=1 Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.
you can't use blackwell gpu with the template, so no 5090 unless you want to do some manual steps to make it compatible
Hi Thanks again for your awesome work. But with your last 2 wan 2.2 workflows I can't find the saved video. It's not in the output directory of my comfyui. I had to deltet the videocombine node and re add it to get it to save. I'm just not sure what I am doing wrong with the default workflow.
well everything is always in the output folder
VERY impressive!!!! 🤔🤗🤷♂️
HI! I've loaded the workflow but I don't see the usual node to toggle sections on and off, if that makes sense. So everything is all purple and disabled. It's probably a user problem.
Yeah. It's missing on mine too. There's no way to use the workflow for me because even the bypass command in the right-click menu has now been removed.
Anyone tell me what version of ComfyUI-VideoHelperSuite they are using? Mine is a fresh install and its saying missing I'm missing VHS_VideoCombine, VHS_LoadVideo, VHS_VideoInfo, RIFE VFI & VHS_SelectImages. I've grep'd in custom_nodes/ComfyUI-VideoHelperSuite and I think this is the plugin, but mine has installed with nightly
never mind a reinstall of the plugins seems to have fixed it...
My question is do you have to put the prompt in for every extension so it can go in the direction you want it to go in. Or does it automatically fill in the prompts based on the last frame, or both? In other words if you leave it on the default and dont change the prompt it will auto make it longer based on the last frame, however if you want to you can change the prompts to guide it, or do you have to put in the prompts from each segment?
I left the example prompts in on purpose so you can see how it works
Where can I get the lora's needed for this to work?
https://huggingface.co/Aitrepreneur/FLX/tree/main
Awesome stuff K! One possible suggestion is to add block swapping to your workflows. I always use your stuff as a jumping off point for a noob like me, but get backed into a corner with my 12gb vram limitation. I always end up finding another workflow that has block swapping which lets some system ram be used more efficiently than vram overflow. I really do love your stuff, but with people like me, something like block swapping could be super helpful.
no need for block swapping since we are using gguf models
A, way better than my approach just to have an final frame extractor and go through the I2V-Workflow once again.
Super. Can you maybe also update the Lora training for 2.2 ?
It was JUST updated, I haven't trained one yet
You are simply amazing! We are very grateful! Thank you so much!
This workflow looks awesome. I am not seeing the download link on this post?
it's here though. https://www.patreon.com/file?h=136696168&m=517615109
Thank you so much ❤️✌🏻✨