Doco here
- dotnet 7.0
- Docker with Qdrant installed as described above
- Azure OpenAI or OpenAI API access
docker-compose -f docker/op-sk-demo/docker-compose.yml up -d
To use OpenAI
- Get your API keys as discussed in the article above
- go to: src/Op.SemanticKernel.Demo.Azure
- Add the user secrets:
dotnet user-secrets set "ApiKey" "The_api_key_you_copied"
To use Azure OpenAI
- Deploy ChatGPT (gpt-35-turbo) and Ada (text-embedding-ada-002) as documented here and copy the deployment names (the name you specified for each one)
- go to: src/Op.SemanticKernel.Demo.Azure
- Add the user secrets:
dotnet user-secrets set "ApiKey" "Key_copied_from_Keys_and_Endpoint"
dotnet user-secrets set "Endpoint" "Endpoint_copied_from_Keys_and_Endpoint"
dotnet user-secrets set "ChatDeploymentName" "Model_deployment_name_of_chat_model"
dotnet user-secrets set "EmbeddingDeploymentName" "Model_deployment_name_of_embedding_model"
- Choose and configure the appropriate starting project ie Op.SemanticKernel.Demo.Azure or Op.SemanticKernel.Demo.OpenAI as per pre-requisites
- Debug/Run
- Profit!
NOTE: On subsequent reruns you can comment out the BuildMemory line as it only needs to be run once to store the document in Qdrant.
- who are you?
- what project did you work on in 2022?