Originally created by @edurenye on GitHub (Feb 7, 2025).
Feature Request
Is your feature request related to a problem? Please describe.
Currently, I can upload JPEG images to the chat and visually capable models will describe it.
But when I try to upload an avif file, it will throw the following error:
Could not detect encoding for /app/backend/data/uploads/file.avif
And then the model will say that it can not interpret images (same visually capable model that worked with the jpeg images).
Describe the solution you'd like
Should just work like with other image formats.
Describe alternatives you've considered
All image formats should work.
Additional context
No additional context.
Originally created by @edurenye on GitHub (Feb 7, 2025).
# Feature Request
**Is your feature request related to a problem? Please describe.**
Currently, I can upload JPEG images to the chat and visually capable models will describe it.
But when I try to upload an avif file, it will throw the following error:
```
Could not detect encoding for /app/backend/data/uploads/file.avif
```
And then the model will say that it can not interpret images (same visually capable model that worked with the jpeg images).
**Describe the solution you'd like**
Should just work like with other image formats.
**Describe alternatives you've considered**
All image formats should work.
**Additional context**
No additional context.
To my knowledge most of the major platforms don't support .avif extensions. JPEGs and PNGs are universally support across all major platforms.
I believe this is likely due to .avif significantly larger file size which would result in a much larger base64 string. So this may actually cause issues if there is a max string size for external model providers.
Testing definitely needed here, but this may be out of our hands unless we actively convert the image to a supported format.
@atgehrhardt commented on GitHub (Feb 8, 2025):
To my knowledge most of the major platforms don't support .avif extensions. JPEGs and PNGs are universally support across all major platforms.
I believe this is likely due to .avif significantly larger file size which would result in a much larger base64 string. So this may actually cause issues if there is a max string size for external model providers.
Testing definitely needed here, but this may be out of our hands unless we actively convert the image to a supported format.
Converting the format could be an interesting solution if the first option is not possible. But in any case a better error message is needed, the current one confuses the users. Something like file format "avif" not supported, please convert the image to one of the supported formats (jpeg, png...)
@edurenye commented on GitHub (Feb 8, 2025):
Converting the format could be an interesting solution if the first option is not possible. But in any case a better error message is needed, the current one confuses the users. Something like file format "avif" not supported, please convert the image to one of the supported formats (jpeg, png...)
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Originally created by @edurenye on GitHub (Feb 7, 2025).
Feature Request
Is your feature request related to a problem? Please describe.
Currently, I can upload JPEG images to the chat and visually capable models will describe it.
But when I try to upload an avif file, it will throw the following error:
And then the model will say that it can not interpret images (same visually capable model that worked with the jpeg images).
Describe the solution you'd like
Should just work like with other image formats.
Describe alternatives you've considered
All image formats should work.
Additional context
No additional context.
@atgehrhardt commented on GitHub (Feb 8, 2025):
To my knowledge most of the major platforms don't support .avif extensions. JPEGs and PNGs are universally support across all major platforms.
I believe this is likely due to .avif significantly larger file size which would result in a much larger base64 string. So this may actually cause issues if there is a max string size for external model providers.
Testing definitely needed here, but this may be out of our hands unless we actively convert the image to a supported format.
@edurenye commented on GitHub (Feb 8, 2025):
Converting the format could be an interesting solution if the first option is not possible. But in any case a better error message is needed, the current one confuses the users. Something like file format "avif" not supported, please convert the image to one of the supported formats (jpeg, png...)
@tjbck commented on GitHub (Feb 18, 2025):
Addressed with
9bd26d6faa