Forum Discussion
Image-to-Image generation using Stable-Diffuson-3.5 Large Model
Has anybody been able to generate an image with 'Image-to-Image' mode of 'Stable-Diffusion 3.5' ( deployed as a serverless api) ?
I tried with text-to-image option (via Python + POST request ) and was able to generate the image but 'image-to-image' option does not seem to work.
Infact even Azure Playground does not list parameters for this option.
But the model information on Azure does list it supports image-to-image mode and accepts an image input.
Any leads on this will be greatly appreciated.
5 Replies
For image-to-image generation using Stable-Diffusion 3.5, ensure that the input image format and parameters align with the model's requirements. Check the API documentation for specific guidelines on image-to-image mode. If the Azure Playground lacks parameters, consider using the Python SDK or REST API directly. If you encounter errors, review the error messages for insights and consider reaching out to Azure support for assistance.
It appears that there are issues with the image-to-image mode of the Stable-Diffusion 3.5 model. I recommend reaching out to the support community or developers for further assistance. Additionally, reviewing the API documentation and checking for any updates or changes that may affect the functionality could be helpful.
It seems you're encountering issues with the image-to-image mode of Stable-Diffusion 3.5. Here are a few suggestions to troubleshoot the problem:
1. **Check Input Format:** Ensure that the input image is in the correct format and meets the requirements of the model. Sometimes, specific formats or resolutions are needed.
2. **Review API Documentation:** Double-check the API documentation for any specific parameters or configurations required for image-to-image generation.
3. **Update Dependencies:** Ensure that all dependencies, including the Stable-Diffusion model and any associated libraries, are up-to-date.
4. **Error Logs:** Check the error logs for any specific error messages that might provide more insight into the issue.
5. **Azure Configuration:** Since you're using Azure, verify that the configuration settings are correctly set up for image-to-image mode.
If the issue persists, consider reaching out to the support community or the developers for further assistance.
I can reproduce the issue. I can generate images with the prompt parameter, but I get the following error trying to use image-to-image:
{"error":{"code":"Invalid input","status":422,"message":"invalid input error","details":[{"type":"Invalid Input","msg":"['image', 'strength', 'mode'] is/are not supported."}]}}
I am getting this error
{"error":{"code":"Bad Request","message":"RAI prompt moderation request failed with status code 424","status":400}}
