reference/code/Fast-dLLM/v1/llada/eval_gsm8k.sh

# Set the environment variables first before running the command.
export HF_ALLOW_CODE_EVAL=1
export HF_DATASETS_TRUST_REMOTE_CODE=true

task=gsm8k
length=256
block_length=32
num_fewshot=5
steps=$((length / block_length))
factor=1.0
model_path='GSAI-ML/LLaDA-8B-Instruct'
# You can change the model path to LLaDA-1.5 by setting model_path='GSAI-ML/LLaDA-1.5'


# baseline
accelerate launch eval_llada.py --tasks ${task} --num_fewshot ${num_fewshot} \
--confirm_run_unsafe_code --model llada_dist \
--model_args model_path=${model_path},gen_length=${length},steps=${length},block_length=${block_length},show_speed=True 


# prefix cache
accelerate launch eval_llada.py --tasks ${task} --num_fewshot ${num_fewshot} \
--confirm_run_unsafe_code --model llada_dist \
--model_args model_path=${model_path},gen_length=${length},steps=${length},block_length=${block_length},use_cache=True,show_speed=True 


# parallel
accelerate launch eval_llada.py --tasks ${task} --num_fewshot ${num_fewshot} \
--confirm_run_unsafe_code --model llada_dist \
--model_args model_path=${model_path},gen_length=${length},steps=${steps},block_length=${block_length},threshold=0.9,show_speed=True

# parallel factor
accelerate launch eval_llada.py --tasks ${task} --num_fewshot ${num_fewshot} \
--confirm_run_unsafe_code --model llada_dist \
--model_args model_path=${model_path},gen_length=${length},steps=${steps},block_length=${block_length},factor=${factor},show_speed=True


# prefix cache+parallel
accelerate launch eval_llada.py --tasks ${task} --num_fewshot ${num_fewshot} \
--confirm_run_unsafe_code --model llada_dist \
--model_args model_path=${model_path},gen_length=${length},steps=${steps},block_length=${block_length},use_cache=True,threshold=0.9,show_speed=True

# dual cache+parallel
accelerate launch eval_llada.py --tasks ${task} --num_fewshot ${num_fewshot} \
--confirm_run_unsafe_code --model llada_dist \
--model_args model_path=${model_path},gen_length=${length},steps=${length},block_length=${block_length},use_cache=True,dual_cache=True,threshold=0.9,show_speed=True

# prefix cache+parallel factor
accelerate launch eval_llada.py --tasks ${task} --num_fewshot ${num_fewshot} \
--confirm_run_unsafe_code --model llada_dist \
--model_args model_path=${model_path},gen_length=${length},steps=${steps},block_length=${block_length},use_cache=True,factor=${factor},show_speed=True

# dual cache+parallel factor
accelerate launch eval_llada.py --tasks ${task} --num_fewshot ${num_fewshot} \
--confirm_run_unsafe_code --model llada_dist \
--model_args model_path=${model_path},gen_length=${length},steps=${length},block_length=${block_length},use_cache=True,dual_cache=True,factor=${factor},show_speed=True