Added how to use commit0 for sampling during STAR training by wenting-zhao · Pull Request #105 · commit-0/commit0
-
Notifications
You must be signed in to change notification settings - Fork 20
Conversation
python examples/star/star.py --model_name_or_path meta-llama/Llama-3.1-8B-Instruct --dataset_name commit0/mbpp -n 10 --output_dir outputs --low_cpu_mem_usage --with_tracking --report_to wandb --iteration 5 --per_device_train_batch_size 1 --gradient_accumulation_steps 8
TODOs:
- Extend STAR to multiple iterations (Done)
- Align the code with the formalism of STaR
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
1 participant
