Qwen2.5-Math/evaluation/sh/eval_sft_cleaned.sh

set -ex

PROMPT_TYPE=qwen25-math-cot
MODEL_NAME_OR_PATH=baidu/Reproducibility/trained_models/trained_3epo_cleaned_torch
OUTPUT_DIR=baidu/Qwen2.5-Math/evaluation/result/ernie_cleaned

SPLIT="test"
NUM_TEST_SAMPLE=-1

# English open datasets
DATA_NAME="gsm8k,math,svamp,asdiv,mawps,carp_en,tabmwp,minerva_math,gaokao2023en,olympiadbench,college_math"
TOKENIZERS_PARALLELISM=false \
python3 -u math_eval.py \
    --model_name_or_path ${MODEL_NAME_OR_PATH} \
    --data_name ${DATA_NAME} \
    --output_dir ${OUTPUT_DIR} \
    --split ${SPLIT} \
    --prompt_type ${PROMPT_TYPE} \
    --num_test_sample ${NUM_TEST_SAMPLE} \
    --seed 0 \
    --temperature 0 \
    --n_sampling 1 \
    --top_p 1 \
    --start 0 \
    --end -1 \
    --use_vllm \
    --save_outputs \
    --overwrite \

# English multiple-choice datasets
DATA_NAME="aqua,sat_math,mmlu_stem"
TOKENIZERS_PARALLELISM=false \
python3 -u math_eval.py \
    --model_name_or_path ${MODEL_NAME_OR_PATH} \
    --data_name ${DATA_NAME} \
    --output_dir ${OUTPUT_DIR} \
    --split ${SPLIT} \
    --prompt_type ${PROMPT_TYPE} \
    --num_test_sample ${NUM_TEST_SAMPLE} \
    --seed 0 \
    --temperature 0 \
    --n_sampling 1 \
    --top_p 1 \
    --start 0 \
    --end -1 \
    --use_vllm \
    --save_outputs \
    --overwrite \
    --num_shots 5

# Chinese gaokao collections
DATA_NAME="gaokao2024_I,gaokao2024_II,gaokao2024_mix,gaokao_math_cloze,gaokao_math_qa"
TOKENIZERS_PARALLELISM=false \
python3 -u math_eval.py \
    --model_name_or_path ${MODEL_NAME_OR_PATH} \
    --data_name ${DATA_NAME} \
    --output_dir ${OUTPUT_DIR} \
    --split ${SPLIT} \
    --prompt_type ${PROMPT_TYPE} \
    --num_test_sample ${NUM_TEST_SAMPLE} \
    --seed 0 \
    --temperature 0 \
    --n_sampling 1 \
    --top_p 1 \
    --start 0 \
    --end -1 \
    --use_vllm \
    --save_outputs \
    --overwrite \
    --adapt_few_shot

# Chinese other datasets
DATA_NAME="cmath,cn_middle_school"
TOKENIZERS_PARALLELISM=false \
python3 -u math_eval.py \
    --model_name_or_path ${MODEL_NAME_OR_PATH} \
    --data_name ${DATA_NAME} \
    --output_dir ${OUTPUT_DIR} \
    --split ${SPLIT} \
    --prompt_type ${PROMPT_TYPE} \
    --num_test_sample ${NUM_TEST_SAMPLE} \
    --seed 0 \
    --temperature 0 \
    --n_sampling 1 \
    --top_p 1 \
    --start 0 \
    --end -1 \
    --use_vllm \
    --save_outputs \
    --overwrite \
    --adapt_few_shot


# English competition datasets
DATA_NAME="aime24,amc23"
TOKENIZERS_PARALLELISM=false \
python3 -u math_eval.py \
    --model_name_or_path ${MODEL_NAME_OR_PATH} \
    --data_name ${DATA_NAME} \
    --output_dir ${OUTPUT_DIR} \
    --split ${SPLIT} \
    --prompt_type ${PROMPT_TYPE} \
    --num_test_sample ${NUM_TEST_SAMPLE} \
    --seed 0 \
    --temperature 0 \
    --n_sampling 1 \
    --top_p 1 \
    --start 0 \
    --end -1 \
    --use_vllm \
    --save_outputs \
    --overwrite \