This is an Llama model that was trained on an instruction dataset that was currated for the purposes of helping people extract meaning, and quickly respond to Google Reviews for their businesses. This model isn't amazing, and thats definitely because the dataset is still lacking sufficient data, so expect this model, as well the training set to improve over time
This llama model was trained 2x faster with
Unsloth and Huggingface's TRL library.
