This workflow is made for who want make long video with low spec PC.
It's a simple way to extract last frame from generated video, then drag it into your I2V workflow to keep generation consistent.
This workflow is made as copypaste , for don't destroy all your workflow.
How to use.
Select all node in group (CTRL+drag area with mouse)
Copy (CTRL+C)
Open your video workflow
Click on an empty area
Paste (CTRL+V)
Connect your VideoCombine-Filenames to "filenames select" node.
All done. Now when you run generation, you get automatic last frame from video.
Tested on wan2.2 and ltx0.9.8
Description
v1.0
FAQ
Comments (18)
(UPDATE: SOLVED) I have a missing node always in Costant: Int-🔬
some reported same issue, workaround is delete the node, and add again.
devilkkw thanks buddy will try now
devilkkw I deleted the node but using add node there was no ComfyUI-logic to choose Int from. Didn't exist even after updating the custom node. What else could I replace this with? What is it actually doing?
SignalMaster the node is for choosing int value.
You can delete it and change the math function in a-1
devilkkw many thanks - now works BEAUTIFULLY and a great idea btw! :)
Fantastic stuff, i wish more people did things like this for example "add upscale nodes to your wf" etc and copy/paste them like this, i have wfs i like a lot but i find it difficult to add new nodes if you don't know exactly what you're doing. About this wf, it worked great, i copy/pasted this on my preferred wf and worked like a charm, the only thing is that i changed the preview node to a save node, thanks a lot for this.
How to connect this to the default template? There is no videocombine-filenames node?
Default workflow Have "CREATE VIDEO" and "SAVE VIDEO" node. Replace this node with "video combine", (remember to configure it). After replace, you are able to paste and use this workflow.
@devilkkw thanks!
I love this extra workflow add-on. I have played with how to auto-add the last frame back into the first chosen image as a loopback for multiple queueing. Here are my efforts (I am only a beginner so be gentle! ;) ):-
If you want to join many video iterations together based on the last frame starting the next queued video, you could try a less manual feedback method - then just queue a few up. This may not be perfect and is experimental!!
Instructions:
1. Copy your starter image into your root output folder, and name it as loop.png
2. In your Load Image dialog in your I2V workflow, choose the loop.png in your output folder as that starter.
3. Add a node to the added 'last frame' workflow. To get this node, search for Image Save from WAS-Suite IO. Link the input of Image Save to the 'Load Video (Path)' IMAGE output
4. Choose the following options in the Image Save dialog:
output path = (delete everything here)
filename prefix = loop
filename delimiter = (delete everything here)
filename number padding = 1
[leave the rest in between]
overwrite_mode = prefix_as_filename
[leave the rest in between]
show_preview = true
Now, when you queue up, each last frame generated will become the new input frame for the next queue position. As I say, this is experimental, but you should get some longer videos. I'm sure this could be further improved but have fun anyway!
nice idea! thank for sharing.
@devilkkw I have since found you need to manually trigger the queue each time but other than that it works well
I'd actually prefer a selection of the last 3-5 frames. Sometimes there is a lot of motion blur in the final frame which makes for a poor starting frame. Up till now I've just done this by hand. How can I configure this to just spit out the last 5 frames? Any ideas.
@Polynechramorph the costant value allow you to select wich frame extract. paste the workflow many time and change the value for each paste.
@devilkkw ok thanks
does not work in my workflow, keep getting errors Loop Detected 126,125,128,129,87,85,
fixed this issue by switching the video saving method to " Video Combine" on my workflow. thanks for sharing this workflow addon
nice to read you have solved.