I just wanted to send a quick message to say thank you. I'm subscribed to your newsletter and the information you share is incredibly valuable. I'm really happy with the service and the amazing work you are doing.
I am on the waiting list for Comfy Cloud and was wondering if you have an estimated launch date. I'm currently using RunPod, and while I'm quite satisfied with it, I would much prefer to give my money to you directly.
The video extend nodes somehow get the wrong video dimensions and all parts of the new video after the first few seconds are zoomed in - how do I fix that?
The mistake is on the extension subgraph. There is no height input. Instead, the input width is connected to both the width and height of the WanAnimateToVideo node. You need to manually add height input to correct that.
Hello. First of all, I would like to thank you for all the work you do. I have a question. Every time I produce a video, I get a zoom effect after the first graph. When the first graph is generated, ( two seconds of vid) the following two seconds are zoomed in. And so on, I have a zoom effect all the way through. How can I fix this? Thank you very much.
wan2.2 animate Scenario: A person is holding a product in front of them. When redrawing the area outside the product, the product will appear severely incompatible. However, with 16fps and a total of 17 frames executed, it can generate a perfect partial image. It is believed that the underlying model supports this.
Cannot find loras / WanAnimate_relight_lora_fp16.safetensors using the link provided in the workflow template. Did it change to something else?
Thank you!
https://huggingface.co/Kijai/WanVideo_comfy/tree/main/LoRAs/Wan22_relight you can find it here
Hi Comfy Team,
I just wanted to send a quick message to say thank you. I'm subscribed to your newsletter and the information you share is incredibly valuable. I'm really happy with the service and the amazing work you are doing.
I am on the waiting list for Comfy Cloud and was wondering if you have an estimated launch date. I'm currently using RunPod, and while I'm quite satisfied with it, I would much prefer to give my money to you directly.
Thanks again for everything!
I'm also curious about it.
We want to develop a service with running ComfyUI in the background, and want to know if we can use it as an end API.
hi how to get the example input image in this blog? thanks! i want to try it.
Is it possible for the Clip to use the new Qwen3VL instead of Qwen 2.5VL as used with the example workflow?
I have tried to run the workflow Qwen-Image-Edit 2509 but running in to black img outputs.
What could be the issue for that? Also is there a workflow where controllnet is implemented?
Thank you
I have 3 nodes missing, named 063b0ee8-c72b-442b-8029-dd3d5f4b6650 and so on. ComfyUI v0.3.64 instead of 0.3.60. Any suggestions?
My comfyUI crashes completely without an error message when processing the node "Video sampling and output.
Is there a way to catch an error?
How to get it work?
This is my hardware:
1 x AMD Ryzen 9 7950X3D
1 x Gigabyte AORUS GeForce RTX 5090 Xtreme WATERFORCE 32G
2 x 48 GB (96 GB) DDR5-5200 Corsair Vengeance
1 x ASUS ProArt | RTX 4080 Super
1 x 1 TB M.2 Samsung 990 PRO | PCIe 4.0 x4
1 x 1 TB M.2 Samsung 990 PRO | PCIe 4.0 x4
1 x Corsair iCUE H100i Elite Capellix XT weiß | 240 mm /2x 120 mm
7 x Be quiet! Pure Wings 3 | 120mm PWM White
1 x Corsair RM1200x SHIFT
Got it to work. Here is the solution:
https://chatgpt.com/s/t_68eac90ca2f4819195be35e1db499a65
Here is the way to there:
https://chatgpt.com/share/68eac955-bf0c-800c-ae0a-29698038f886
ran both bat-files, still goes straight to cpu (rtx5090), only wan animate does this, any idea?
The video extend nodes somehow get the wrong video dimensions and all parts of the new video after the first few seconds are zoomed in - how do I fix that?
The mistake is on the extension subgraph. There is no height input. Instead, the input width is connected to both the width and height of the WanAnimateToVideo node. You need to manually add height input to correct that.
I have the same problem, I am interested in its resolution.
Every time when i try to connect "face video" it will generate fully black video.
Hello. First of all, I would like to thank you for all the work you do. I have a question. Every time I produce a video, I get a zoom effect after the first graph. When the first graph is generated, ( two seconds of vid) the following two seconds are zoomed in. And so on, I have a zoom effect all the way through. How can I fix this? Thank you very much.
Hello. Every time generates black video. Only black frames
wan2.2 animate Scenario: A person is holding a product in front of them. When redrawing the area outside the product, the product will appear severely incompatible. However, with 16fps and a total of 17 frames executed, it can generate a perfect partial image. It is believed that the underlying model supports this.
lora key not loaded
relight_lora doesn't seem to be working.
Thank you!
WanAnimate_relight_lora_fp16.safetensors
download from
https://huggingface.co/Kijai/WanVideo_comfy/tree/main/LoRAs/Wan22_relight
Cannot find loras / WanAnimate_relight_lora_fp16.safetensors using the link provided in the workflow template. Did it change to something else?
Thank you!
Cannot find loras / WanAnimate_relight_lora_fp16.safetensors using the link provided in the workflow template. Did it change to something else?
Thank you!