Closed Tangroo closed 1 year ago
Are you able to use the endpoint in the Azure OpenAI playground?
Based on the provided image, it doesn't look like your REACT_APP_BACKEND_URI
is updated to your deployed instance, it is still pointing to localhost
.
cc - @glahaye
I try to deploy the copilot chat app, use AzureOpenAI, and try to chat after deployment, but when I send a message to it, an error "Error: Invalid request: The request is not valid, HTTP status: 404. Details: The API deployment for this resource does not exist. If you created the deployment within the last 5 minutes, please wait a moment and try again." will be reported. What is the reason for this?
Hey man, did you create the deployments for the models that you will need to run copilot? if not then go ahead and do that, that should solve the problem, the names for deployments should be the same as the model names...
Are you able to use the endpoint in the Azure OpenAI playground?
yes, I used
Based on the provided image, it doesn't look like your
REACT_APP_BACKEND_URI
is updated to your deployed instance, it is still pointing tolocalhost
.cc - @glahaye
Based on the provided image, it doesn't look like your
REACT_APP_BACKEND_URI
is updated to your deployed instance, it is still pointing tolocalhost
.cc - @glahaye
I deployed it locally, both webapi and webapp were deployed locally
I try to deploy the copilot chat app, use AzureOpenAI, and try to chat after deployment, but when I send a message to it, an error "Error: Invalid request: The request is not valid, HTTP status: 404. Details: The API deployment for this resource does not exist. If you created the deployment within the last 5 minutes, please wait a moment and try again." will be reported. What is the reason for this?
Hey man, did you create the deployments for the models that you will need to run copilot? if not then go ahead and do that, that should solve the problem, the names for deployments should be the same as the model names...
Excuse me, what do you mean by model? I just configured the environment locally first, then directly deployed the webapi and webapp of copilot chat locally, then created an authentication app on Azure according to the README for authentication, then opened Azure OpenAI, and configured the key and endpoint of AzureOpenAI to webapi, all follow the steps on the README. Are there any other models that need to be deployed?
First, do you have an OpenAI account WITH AN API key? If you don't have an API key, you won't be able access OpenAI resources through the API's offered at their endpoint.
Then, assuming you do have an account with its associated API key, does that account have the models requested in the appsettings,json file deployed to it?
Also, if you are using OpenAI (as opposed to Azure OpenAI), you should not reference gpt-35-turbo but rather gpt-3.5-turbo (notice the '.' between 3 and 5) in your appsettings.json file for the completion and planner models.
sorry, it's my fault, I didn't deploy the semantic kernel to the Azure. I only deployed the copilot-chat app.
I can chat with it now, but I feel that the quality of his answer is not as high as chatgpt3.5. What is the reason? I used AzureOpenAI, gpt-35-turbo
Poor response quality could be due to the prompts used within the chat skill.
@TaoChenOSU ? @adrianwyatt ?
I can chat with it now, but I feel that the quality of his answer is not as high as chatgpt3.5. What is the reason? I used AzureOpenAI, gpt-35-turbo
Experimenting with different prompts would be the best way to improve the responses. It's currently biased towards a long running chat scenario which may not be best for your scenario.
The current way to modify the prompt used is to tweak the values in appsettings.json. You may have to do this and restart the aspnet app, not sure if changes will be picked up while it's running.
I'm going to close this for now as it seems the 404 issue has been resolved, feel free to reopen if not and/or to open a new issue if you aren't able to get the prompts to behave the way you want them to.
sorry, it's my fault, I didn't deploy the semantic kernel to the Azure. I only deployed the copilot-chat app.
How did you do this?
I try to deploy the copilot chat app, use AzureOpenAI, and try to chat after deployment, but when I send a message to it, an error "Error: Invalid request: The request is not valid, HTTP status: 404. Details: The API deployment for this resource does not exist. If you created the deployment within the last 5 minutes, please wait a moment and try again." will be reported. What is the reason for this?