1 comment

[ 3.4 ms ] story [ 9.3 ms ] thread
This project provides high-level functionality on top of the popular LLMs, allowing you to decouple the models and switch between them based on your use case.

The available functionality:

- Chatbot: define the provider as chatGPT or llama chat.

- Evaluation: Run evaluation across multiple models by sending a query and array of target answers. This will call the models in the background, generate the vector, and compare the distances. This supports cohere, openai, replicate, and sage maker models.

- Semantic Search: Apply search beyond the keywords using vectors.

- Direct models: directly access the model providers like Sage maker llama or hugging face.

- Offline model loader: under development.

The micro service published to docker hub:

```

docker pull intellinode/intelliserver:latest

docker run -p 80:80 -e API_KEY=$API_KEY -e ADMIN_KEY=$ADMIN_KEY intellinode/intelliserver:latest

```

I am working to add more features to the micro-service, Let me know which functions you think going to be useful.