Skip to content

Latest commit

 

History

History
27 lines (15 loc) · 1.13 KB

README.md

File metadata and controls

27 lines (15 loc) · 1.13 KB

NOTE: This repository is a work in progress. Changes and updates may occur as the project evolves.

ARC-TTT

Simplified and Stripped-Down ARC-Test-Time-Training (ARC-TTT). This implementation uses Hugging Face and unsloth, based on the original MARC repository, which used torchtune. Credits to MARC.

Env Setup:

Create your conda environment for training with unsloth as explained here.

Run Arc TTT

Does data augmentation on the training tasks and fine-tunes adapter per arc task and runs evaluation on the test task. Does not utilize a fine-tuned base model yet.

python ttt_unsloth.py

Original implementation uses a fine-tuned base model. In this repo we fine-tune task adapter from scratch of a regular pre-trained model that is not yet fine-tuned on the ARC tasks.

Results

TTT performance over 20 tasks compared to untrained base model:

alt text