​
Anyscale Endpoints Handler

This integration allows you to connect your AnyscaleEndpoints models with MindsDB as AI tables through the AnyscaleEndpoints ML engine. The implementation is based on the engine for the OpenAI API, as Anyscale conforms to it. There are a few notable differences, though:
  1. The supported models are completely different.
    1. All models supported by AnyscaleEndpoints are open source. A full list can be found here for inference-only, under section “Supported models”.
    2. For fine-tuning, not every model is supported. You can find a list here, under section “Fine Tuning - Supported models”.
    3. Please check both lists regularly, as they are subject to change. If you try to fine-tune a model that is not supported, you will get a warning and subsequently an error from the Anyscale endpoint.
  2. This integration only offers chat-based text completion models, either for “normal” text or specialized for code.
  3. When providing a description, this integration returns the respective HuggingFace model card.
  4. Fine-tuning requires that your dataset complies with the chat format. That is, each row should contain a context and a role. The context is the text that is the message in the chat, and the role is who authored it (system, user or assistant, where the last one is the model). For more information, please check the fine tuning guide in the Anyscale Endpoints docs.

​
Inference example

Here is an example to produce sentiment analysis predictions on some text: 
CREATE ML_ENGINE anyscale_endpoints FROM anyscale_endpoints USING api_key = 'your-api-key';
    
CREATE MODEL anyscale_endpoints_sentiment_model
PREDICT sentiment
USING
engine = 'anyscale_endpoints',
model_name = 'mistralai/Mistral-7B-Instruct-v0.1',
prompt_template = 'Classify the sentiment of the following text as one of `positive`, `neutral` or `negative`: {{text}}';
    
SELECT sentiment 
FROM anyscale_endpoints_sentiment_model
WHERE text = 'I love machine learning!';