v0.3 - New 4bit quantization support

This release:

• Updated all code to latest
• Supports new 4bit models
• Uses Alpaca 7b model from https://huggingface.co/ozcur/alpaca-native-4bit by default

v0.1

First downloadable release.