metadata
license: mit
This repo contains a low-rank adapter for LLaMA-7b finetuned on Ntropy proprietary dataset (consumer financial transactions).
This version of the weights was trained with the following hyperparameters:
- Base Model: decapoda-research/llama-7b-hf
- Epochs: 10 (load from best epoch)
- Batch size: 32
- Cutoff length: 1024
- Learning rate: 3e-4
- Lora r: 16
- Lora target modules: q_proj, k_proj, v_proj, o_proj
Instructions for running it can be found at https://github.com/ntropy-network/enrichment_models/blob/main/notebooks/llama.ipynb