metadata

library_name: peft
base_model: TinyLlama/TinyLlama-1.1B-Chat-v0.4
license: apache-2.0

This model was trained by me on the Alpaca dataset.

Benchmarks

Tests were carried out on this model and the original model for the purpose of accurate data.

boolq - 59.05 (original - 57.31)

truthfulqa_mc1 - 23.79 (original - 23.01)

ARC_e - 52.27 (original - 54.04)

WinoGrande - 58.3 (original - 58.8)