-
Notifications
You must be signed in to change notification settings - Fork 4.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[Issue Report] Model Addition Failure: Local Models Not Displayed in System-Model Settings Search Results #4450
Comments
It sounds like the core issue you're experiencing is that newly deployed models aren't showing up in the System-Model Settings search results as expected. This can be addressed by ensuring that the models are properly registered and recognized by the system immediately after deployment, without requiring the creation of a new application as a workaround. To resolve this, you should utilize the Here's a brief overview of steps to integrate this solution:
For example, if you've added or updated models related to text generation and reranking, you would call updateModelList(ModelTypeEnum.textGeneration);
updateModelList(ModelTypeEnum.rerank); This approach is particularly useful in environments where models are frequently updated or deployed, as it automates the visibility of these models in the System-Model Settings. Make sure that the models are correctly categorized under specific model types as defined by the system (e.g., For detailed implementation, refer to the If you encounter any further issues or have questions about this process, feel free to ask.
|
现在是否支持自定义ReRank模型? |
@dosu irrelevant answer |
@wanshichenguang can, Use the same interface protocol |
Self Checks
Dify version
0.6.5
Cloud or Self Hosted
Self Hosted (Docker)
Steps to reproduce
During the following operations, I encountered an issue: I successfully deployed ReRank, Text Embedding, and Chat models on the server using XinFerence and OpenAI-API-compatible methods. The deployment process indicated that everything was completed successfully, suggesting that these models should be ready for use.
However, when attempting to locate these newly added local models within the System-Model Settings, they were not appearing in the search results. This situation hinders the intended workflow of directly accessing or managing these models from this interface.
Reproduction Steps:
Deploy the ReRank, Text Embedding, and Chat models using XinFerence and OpenAI-API-compatible tools.
After confirming successful deployment, access the system administration interface.
Navigate to the System-Model Settings section and attempt to search for the newly deployed models.
Expected Result:
Upon entering the model name in the search bar of the System-Model Settings, the relevant local models should promptly appear in the search results.
Actual Result:
The models were not retrieved; the search results were empty.
Temporary Solution:
Currently, to make these models visible and selectable in the System-Model Settings, additional steps are required:
Create a new application first.
Within this application, manually select the chat model that was previously deployed.
Only after launching the application do these models become visible within the System-Model Settings interface.
Desired Resolution:
Please investigate and fix this issue to ensure that future deployments of local models are instantly and accurately reflected in the System-Model Settings search results without necessitating further manual intervention.
✔️ Expected Behavior
Upon entering the model name in the search bar of the System-Model Settings, the relevant local models should promptly appear in the search results.
❌ Actual Behavior
The models were not retrieved; the search results were empty.
The text was updated successfully, but these errors were encountered: