The chatBot's image upload model, using the gpt4o model, cannot recognize images. It was functioning properly before the upgrade.
Self Checks
- [X] This is only for bug report, if you would like to ask a question, please head to Discussions.
- [X] I have searched for existing issues search for existing issues, including closed ones.
- [X] I confirm that I am using English to submit this report (我已阅读并同意 Language Policy).
- [X] [FOR CHINESE USERS] 请务必使用英文提交 Issue,否则会被关闭。谢谢!:)
- [X] Please do not modify this template :) and fill in all the required fields.
Dify version
0.10.2~0.11.0
Cloud or Self Hosted
Self Hosted (Docker)
Steps to reproduce
The chatBot's image upload model, using the gpt4o model, cannot recognize images. It was functioning properly before the upgrade.
✔️ Expected Behavior
I hope the model will be able to recognize images normally.
❌ Actual Behavior
cannot recognize images
I found some similar discussions and issues related to image processing problems in Dify:
- Unable to open image: /files/tools - This is an unsolved discussion that might be related to your issue.
- I can't upload the image on dify workflow via api - Another unsolved discussion that could be relevant.
- GPT-4V using bug, There is no image input but output. The image link was successfully uploaded, but an error message was reported that the URL does not exist - This issue is closed, and it was suggested to check the Docker Nginx configuration to ensure the necessary block for file accessing is present, as the Nginx config was recently updated to support this.
The recent upgrade of Dify replaced the standalone image upload feature with a more comprehensive "File Upload" feature, which might have affected the image recognition functionality. The image upload option in Chatflow's "Features" has been removed and replaced by the new "File Upload" feature, where users can select the image file type [1][2].
You might want to check your Nginx configuration and ensure it aligns with the latest updates.
To continue talking to Dosu, mention @dosu.
I tested this feature on our Cloud service, and it looks fine.
I continued testing and found that images pasted using CTRL+V cannot be recognized, while manually uploaded ones can be recognized.
I continued testing and found that images pasted using CTRL+V cannot be recognized, while manually uploaded ones can be recognized.
After testing, it seems that the first pasted image cannot be recognized, but the second one works normally.
Hi, @GeorgeGalway. I'm Dosu, and I'm helping the Dify team manage their backlog. I'm marking this issue as stale.
Issue Summary
- The image upload feature stopped recognizing images after upgrading to the gpt4o model in a self-hosted Docker environment.
- I suggested checking the Nginx configuration and noted a change to a "File Upload" feature.
- Laipz8200 reported no issues with the feature on their Cloud service.
- You found that images pasted using CTRL+V are initially not recognized, but subsequent pastes work, while manual uploads are fine.
Next Steps
- Please confirm if this issue is still relevant with the latest version of the Dify repository. If so, you can keep the discussion open by commenting here.
- Otherwise, this issue will be automatically closed in 15 days.
Thank you for your understanding and contribution!