codellama_7b_generate_baseline_70.sh

#!/bin/sh
eval "$(conda shell.bash hook)"

PROJECT_ROOT=$(cd $(dirname $0); cd ../; pwd -P)
echo "Project root path: $PROJECT_ROOT"

conda activate replay-env


# shellcheck disable=SC2112
#          --prompt_style infill \
#function codellama_generate() {
#    history_category_flag=1
#    style=instruct
#    for device_id in 0 1 2 3 4 5 6; do
#      CUDA_VISIBLE_DEVICES=$device_id nohup \
#      python -u codellama_generate.py \
#          --prompt_style $style \
#          --subject_model_id codellama/CodeLlama-7b-Instruct-hf \
#          --history_category $history_category_flag \
#          --bugs_meta_data_file $PROJECT_ROOT/dataset/bugs_meta_data.json \
#          --bugs_description_file $PROJECT_ROOT/dataset/github_issue/bugs_description.json \
#          --history_data_path $PROJECT_ROOT/dataset/history_blame \
#          --result_output_path $PROJECT_ROOT/model_inference_$((device_id+7)) \
#          --has_nucleus_sampling 1 \
#          --is_buggy_line_labeled 0 > log/baseline70/codellama_7b_${style}_${history_category_flag}_$((device_id+7)).log &
#      echo "--------------------history_category_flag: $history_category_flag---------------------"
#    done
#}
#function codellama_generate() {
#    history_category_flag=1
#    style=instruct
#    for device_id in 0 1 2 3 4 5 6; do
#      CUDA_VISIBLE_DEVICES=$device_id nohup \
#      python -u codellama_generate.py \
#          --prompt_style $style \
#          --subject_model_id codellama/CodeLlama-7b-Instruct-hf \
#          --history_category $history_category_flag \
#          --bugs_meta_data_file $PROJECT_ROOT/dataset/bugs_meta_data.json \
#          --bugs_description_file $PROJECT_ROOT/dataset/github_issue/bugs_description.json \
#          --history_data_path $PROJECT_ROOT/dataset/history_blame \
#          --result_output_path $PROJECT_ROOT/model_inference_label_${history_category_flag}_${device_id} \
#          --has_nucleus_sampling 1 \
#          --is_buggy_line_labeled 1 > log/baseline70/${style}/codellama_7b_${style}_${history_category_flag}_${device_id}.log &
#      echo "--------------------history_category_flag: $history_category_flag---------------------"
#    done
#}

function codellama_generate() {
    history_category_flag=1
    style=infill
    for device_id in 0 1 2 3 4 5 6; do
      CUDA_VISIBLE_DEVICES=$device_id nohup \
      python -u codellama_generate.py \
          --prompt_style $style \
          --subject_model_id codellama/CodeLlama-7b-Instruct-hf \
          --history_category $history_category_flag \
          --bugs_meta_data_file $PROJECT_ROOT/dataset/bugs_meta_data.json \
          --bugs_description_file $PROJECT_ROOT/dataset/github_issue/bugs_description.json \
          --history_data_path $PROJECT_ROOT/dataset/history_blame \
          --result_output_path $PROJECT_ROOT/model_inference_infill_${history_category_flag}_${device_id} \
          --has_nucleus_sampling 1 \
          --is_buggy_line_labeled 0 > log/baseline70/${style}/codellama_7b_${style}_${history_category_flag}_${device_id}.log &
      echo "--------------------history_category_flag: $history_category_flag---------------------"
    done
}

codellama_generate
conda deactivate