[Question]: azure openai menu not appearing #2484
Replies: 13 comments 19 replies
-
|
I had to re-setup everything and I can t get openrouter to appear for anything.. I feel your pain man.. I dont know if new updates to repo in past day or 2 messed things up |
Beta Was this translation helpful? Give feedback.
-
|
@qcpsigmund @bpawnzZ You likely have an invalid custom config file. That would mean that your custom configurations would not be used. I'm considering throwing errors for invalid configurations, since that would prevent the app from running at all and making the issues more obvious. That said, please check and share logs so we can help. If using docker, you can find them in the ./logs directory of project root. |
Beta Was this translation helpful? Give feedback.
-
i see this error in my logs, so i assume a valid azure openapi key has to be entered with the respective models before the endpoint will be enabled? |
Beta Was this translation helpful? Give feedback.
-
|
Not a valid key, but a valid configuration. Your issue is that "interface" is not recognized as a valid key or field in the configuration. This shouldn't be an issue if you are on the latest version, which there might be an issue with how you're updating. See the docs here on how to update docker: https://docs.librechat.ai/install/installation/docker_compose_install.html#updating-librechat but you can resolve this easily by removing this from your config file: # Custom nterface configuration
interface:
# Privacy policy settings
privacyPolicy:
externalUrl: 'https://librechat.ai/privacy-policy'
openNewTab: true
# Terms of service
termsOfService:
externalUrl: 'https://librechat.ai/tos'
openNewTab: truethen you might have a new error if there's still issues with the file. |
Beta Was this translation helpful? Give feedback.
-
|
i did the update and removed the interface in the yaml config. the error logs disappeared but I am getting this in debug logs now |
Beta Was this translation helpful? Give feedback.
-
|
Use https://yamlchecker.com/ on your config |
Beta Was this translation helpful? Give feedback.
-
|
this is my current yaml after using yamlchecker. still getting 2024-04-22T13:18:51.453Z info: Custom config file missing or YAML format invalid. |
Beta Was this translation helpful? Give feedback.
-
|
Perform an update as linked to earlier and then share your logs |
Beta Was this translation helpful? Give feedback.
-
|
i have removed the entire installation, git clone the current repo and ran docker compose pull. updated the yaml and env file values. .env file librechat.yaml file |
Beta Was this translation helpful? Give feedback.
-
|
likely missing the override file, did you do this step? https://docs.librechat.ai/install/configuration/custom_config.html#docker-setup |
Beta Was this translation helpful? Give feedback.
-
|
this should be the configuration in the override file? |
Beta Was this translation helpful? Give feedback.
-
|
great i see it now, however it shows the API error
|
Beta Was this translation helpful? Give feedback.
-
|
if i use the following code from azure, i am able to authenticate this is my current librechat.yaml file |
Beta Was this translation helpful? Give feedback.
-
|
just to confirm the instance name is supposed to be Azure-OpenAI-RG or Azure-OpenAI-QCP? i tried both but still got the same error |
Beta Was this translation helpful? Give feedback.
-
|
what do your deployments look like?
|
Beta Was this translation helpful? Give feedback.
-
|
these are my current 2 deployments. |
Beta Was this translation helpful? Give feedback.
-
|
what is the name of your deployments? from what I can tell, it's the model name in your screenshot? in which case, you did not configure the yaml correctly. The point of this config is to enable you to name your deployment whatever you want, and then map it to an actual model name, which the app recognizes for context size. Also, you can use multiple deployments this way. With the legacy setup, you are forced to use the deployment name in the model list, or only use one deployment for all models (so model value is basically useless) |
Beta Was this translation helpful? Give feedback.
-
|
i think i get what you mean. will try and update again! |
Beta Was this translation helpful? Give feedback.
-
|
weirdly if i use the legacy setup, it works with the the 3.5/4 model. however, when using the gpt4 model and uploading an image, it shows error 400 invalid content type. image_url is only support by certain models |
Beta Was this translation helpful? Give feedback.
-
|
i have tried deploying the vision model as gpt4 does base model does not support image input. does my config has any errors? the previous gpt4 base model in us east 2 region was working with librechat with another config
working config with gpt4 |
Beta Was this translation helpful? Give feedback.
-
|
gpt-4-vision-preview needs to be an available model. with the azure config, it can be done easily like this: models:
gpt-4-vision-preview:
deploymentName: "your-gpt-4-vision-deployment-name" |
Beta Was this translation helpful? Give feedback.
-
|
without it, gpt-4-vision-preview must be available, however you configure your model list |
Beta Was this translation helpful? Give feedback.
-
|
amazing now the yaml file works! let me try to deploy the gpt3.5 model as well and hopefully it still works. thank you for your help! |
Beta Was this translation helpful? Give feedback.
-
|
you're welcome! |
Beta Was this translation helpful? Give feedback.
-
|
hi @danny-avila another issue i face was integrating the assistants api into the yaml file(in eastus2 region and 1106 model). this is my current config. can you have a look? i am also unable to upload files other than images for the models to decipher. is there any config im missing in the models? |
Beta Was this translation helpful? Give feedback.
-
|
Maybe I can help, as I had the exact same issue with no assistants or plugin UI showing. This is my config, in case it helps (you want to remove retrieval from the assistants capabilities as it's not supported in azure), also you need to add a text-embeddings model in .env so that it can upload files to it. `endpoints: azureOpenAI: My .env (please note that Dall-E doesn't seem to work for me, it's the only thing missing...) `RAG_AZURE_OPENAI_API_KEY=apiKey DALLE3_AZURE_API_VERSION=2024-02-15-preview |
Beta Was this translation helpful? Give feedback.
-
|
hi @maxesse so for RAG do i have to deploy any models or azure services such as azure search to get the api key?i would have to deploy a text-embedding model in my resource group? for text embedding model configuration, it will be done in .env instead of librechat.yaml file? |
Beta Was this translation helpful? Give feedback.
-
|
Yeah, just deploy a text-embeddings model in your openAI deployment. I used ada-002 because that's what's available in sweden - in the US i think you can use large-3, which is nicer. |
Beta Was this translation helpful? Give feedback.
-
|
@danny-avila would you be able to advise on RAG as well? i have deployed the text-embedding-ada-002 model in my azure in the same region as the vision model(japeast) and i get this error. this is my current librechat.yaml file. the vision model seems to still work for chat and images but the gpt 4 model in the US is not working with this error .env file |
Beta Was this translation helpful? Give feedback.
-
What is your question?
i am unable to select the azure open ai endpoint with my .env and yaml file config
More Details
What is the main subject of your question?
Endpoints
Screenshots
Code of Conduct
Beta Was this translation helpful? Give feedback.
All reactions