Using Conda is recommended for managing dependencies and creating the virtual environment.
- Ensure you have Conda installed on your system.
- Create a new Conda environment:
conda create --name myenv python=3.x
- Activate the environment:
conda activate myenv
- Install required packages:
# If you have a requirements file, use: pip install -r requirements.txt
The script uses the following default values. You can modify these as needed:
DATASET_FILE
: Path to your dataset file.PROJECT_NAME
: Name of your project.MODELS_LIST_DIR
: Directory containing model configurations.JUDGE_TEMPLATES_FILE
: Path to the judge templates file.WOKE_TEMPLATES_FILE
: Path to the woke templates file.SCRIPTS_DIR
: Directory of your scripts.STAGE_6_FILE
: Optional stage 6 file.INSERTION_POINT
: Default insertion point.CUDA_VISIBLE_DEVICES
: GPUs allocated for the task.RAY_TMPDIR
: Temporary directory for Ray.
#!/bin/bash
# Set your default values here
DATASET_FILE="$(pwd)/data/HEx-PHI/category_1.csv"
PROJECT_NAME="Demo"
MODELS_LIST_DIR="$(pwd)/configs/open_models.txt"
JUDGE_TEMPLATES_FILE="$(pwd)/configs/judge_prompt.jsonl"
WOKE_TEMPLATES_FILE="$(pwd)/configs/woke_templates.json"
SCRIPTS_DIR="$(pwd)/scripts"
STAGE_6_FILE=""
INSERTION_POINT=0
CUDA_VISIBLE_DEVICES="0,1,2,3"
RAY_TMPDIR="$HOME/tmp_ray"
# Run the main pipeline script with arguments
./main_pipeline.sh \
--dataset_file "${1:-$DATASET_FILE}" \
--project_name "${2:-$PROJECT_NAME}" \
--models_list_dir "${3:-$MODELS_LIST_DIR}" \
--judge_templates_file "${4:-$JUDGE_TEMPLATES_FILE}" \
--woke_templates_file "${5:-$WOKE_TEMPLATES_FILE}" \
--scripts_dir "${6:-$SCRIPTS_DIR}" \
--stage_6_file "${7:-$STAGE_6_FILE}" \
--insertion_point "${8:-$INSERTION_POINT}" \
--cuda_visible_devices "${9:-$CUDA_VISIBLE_DEVICES}" \
--ray_tmpdir "${10:-$RAY_TMPDIR}"
To execute the pipeline, navigate to the directory containing setup.sh
and run:
./setup.sh
This documentation provides clear, step-by-step instructions for setting up the environment and running the code. Follow these instructions to quickly get your pipeline up and runnin