EarlyStopping

This repo has all the code to run the Chain-of-Embedding code and run clustering algorithms for BoN

First run main.py to run inference on the dataset and save the COE scores and hidden_states

Then run calc_reward to save the rewards

Then run any of the filter bon, 2inference or good turing for different experiments.

Name		Name	Last commit message	Last commit date
Latest commit History 8 Commits
.vscode		.vscode
Data		Data
Evaluation		Evaluation
Scripts		Scripts
par-seq		par-seq
reward_model_config		reward_model_config
README.md		README.md
amir-test.py		amir-test.py
armorm.py		armorm.py
calculate_reward_armor.py		calculate_reward_armor.py
calculate_reward_qwen.py		calculate_reward_qwen.py
calculate_reward_skywork.py		calculate_reward_skywork.py
config_pool.py		config_pool.py
create-dataset.py		create-dataset.py
eval_bon.py		eval_bon.py
filter-bon.py		filter-bon.py
good_turing.py		good_turing.py
inference.py		inference.py
inference_batch.py		inference_batch.py
inference_batch_vlllm.py		inference_batch_vlllm.py
inference_self_correct.py		inference_self_correct.py
main.py		main.py
prompt_pool.py		prompt_pool.py
reward_model.py		reward_model.py
score.py		score.py
two_inference.py		two_inference.py

Provide feedback