Video tutorial on how to use the tool
Points to considers
1. Please make sure Automatic 1111 is running and openpose extension is installed
2. If you are using localhost url, please don't forget to enable --api launch flag for automatic 1111
3. Once generate button is click please wait for the images to be generated
4. Currently tested for colab and localhost only but should run for runpod instance also
5. Prompt generation framework
Checkout github page for the code
https://github.com/dinilvalsan/random-pose-generator
Checkout https://controlnetposes.com/ for tons of free poses!
Description
FAQ
Comments (7)
How do you set up high res fix and scaling on this?
@SpicyBell From my small experience in creating large batches, only 40 - 50% image are good for usable quality, as it would be waste of resource to run a expensive upscale with all the batch. A suggestion is to short list items after generation and upscale each one manually using gigapixel or automatic upscalers. Also i dont remember it were it was, may in reddit stablediffusion subreddit, a user had posted batch upscaler script. If your comment gets enough upvotes i will try to add a checkbox for that in next release :)
an awesome batch program! i don't see the multi control net option, however. am i just missing something? or is there a possibility you could update it for that?
also, a low vram checkbox would be awesome too : )
I love this concept! But can you use loras with the prompts?
This used to work for me but now it crashes whenever I try fetch Samplers (but I can fetch ControlNet models without issue). Does this still work or should I use something else?
Did you ever find a solution for this? I only just discovered this little program and it's the exact problem I'm having.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.