Model Card for TarsMeta
Model Details
TarsMeta is FT on MetaMathQA dataset for better word math problem solving capabilities
Training Data
epochs 4
rows 39K
lr 2e-4
Training Procedure
autotrain-advanced
Preprocessing
dataset was converted in chatML format
from adapters import AutoAdapterModel model = AutoAdapterModel.from_pretrained("undefined") model.load_adapter("xriminact/TarsMeta", set_active=True)