Lambda function

[stefan-novak-brt]

Here's an initial work of deploying a RetrievalQA inside of a Lambda function. The approach here is to load the sample dataset into a Chroma database and persist it during the image build phase. When the Lambda runs, it'll be able to open the database on-the-fly.

Steps to run:

1. Download ggml-alpaca-7b-q4.bin to the models/ directory.
2. Run make image

Once the Docker image can be built, we can add a Terraform config to deploy the Lambda function to AWS. We can hook up the Lambda to the AWS API Gateway so that a web app can call it via a REST API.

[stefan-novak-brt]

When lambda.py runs during the build phase, the following error is occurring:

 > [8/8] RUN python3 lambda.py:
#0 3.412 llama.cpp: loading model from models/ggml-alpaca-7b-q4.bin
#0 81.88 terminate called after throwing an instance of 'std::runtime_error'
#0 81.88   what():  unexpectedly reached end of file
#0 95.02 Aborted (core dumped)