cd opencompass
pip install -e .


# For HumanEval evaluation, install the additional dependency:
git clone https://github.com/open-compass/human-eval.git
cd human-eval && pip install -e .
cd ..


# For Math evaluation, pip install the additional dependency:
pip install math_verify latex2sympy2_extended


# LLaDA-8B-Base
python run.py examples/llada_base_gen_gsm8k_length256_block256.py -w outputs/llada_base_gsm8k_length256_block256
python run.py examples/llada_base_gen_math_length256_block256.py -w outputs/llada_base_math_length256_block256
python run.py examples/llada_base_gen_humaneval_length256_block256.py -w outputs/llada_base_humaneval_length256_block256
python run.py examples/llada_base_gen_mbpp_length256_block256.py -w outputs/llada_base_mbpp_length256_block256
python run.py examples/llada_base_gen_bbh_length256_block256.py -w outputs/llada_base_bbh_length256_block256



# LLaDA-8B-Instruct
python run.py examples/llada_instruct_gen_mmlu_length3_block3.py -w outputs/llada_instruct_mmlu_length3_block3
python run.py examples/llada_instruct_gen_mmlupro_length256_block256.py -w outputs/llada_instruct_mmlupro_length256_block256
python run.py examples/llada_instruct_gen_hellaswag_length3_block3.py -w outputs/llada_instruct_hellaswag_length3_block3
python run.py examples/llada_instruct_gen_arcc_length512_block512.py -w outputs/llada_instruct_arcc_length512_block512
python run.py examples/llada_instruct_gen_gsm8k_length512_block512_confidence.py -w outputs/llada_instruct_gsm8k_length512_block512_confidence
python run.py examples/llada_instruct_gen_math_length512_block512_confidence.py -w outputs/llada_instruct_math_length512_block512_confidence
python run.py examples/llada_instruct_gen_gpqa_length64_block64_confidence.py -w outputs/llada_instruct_gen_gpqa_length64_block64_confidence
python run.py examples/llada_instruct_gen_humaneval_length512_block512_logits.py -w outputs/llada_instruct_gen_humaneval_length512_block512_logits
python run.py examples/llada_instruct_gen_mbpp_length256_block256_confidence.py -w outputs/llada_instruct_gen_mbpp_length256_block256_confidence
python run.py examples/llada_instruct_gen_ifeval_length512_block512_confidence.py -w outputs/llada_instruct_gen_ifeval_length512_block512_confidence


python run.py examples/llada_instruct_gen_gsm8k_length256_block8.py -w outputs/llada_instruct_gen_gsm8k_length256_block8
python run.py examples/llada_instruct_gen_math_length512_block64.py -w outputs/llada_instruct_gen_math_length512_block64



# LLaDA 1.5
python run.py examples/llada_1p5_gen_gsm8k_length256_block16_confidence.py -w outputs/llada_1p5_gen_gsm8k_length256_block16_confidence
python run.py examples/llada_1p5_gen_math_length1024_block128_confidence.py -w outputs/llada_1p5_gen_math_length1024_block128_confidence
python run.py examples/llada_1p5_gen_gpqa_length256_block16.py -w outputs/llada_1p5_gen_gpqa_length256_block16
python run.py examples/llada_1p5_gen_humaneval_length512_block32_confidence.py -w outputs/llada_1p5_gen_humaneval_length512_block32_confidence
python run.py examples/llada_1p5_gen_mbpp_length512_block32_confidence.py -w outputs/llada_1p5_gen_mbpp_length512_block32_confidence
python run.py examples/llada_1p5_gen_ifeval_length256_block16_confidence.py -w outputs/llada_1p5_gen_ifeval_length256_block16_confidence