Add auto data download for all scripts

OptimalScale · Aug 7, 2023 · dff1f92 · dff1f92
1 parent db91533
commit dff1f92
Show file tree

Hide file tree

Showing 6 changed files with 23 additions and 0 deletions.
diff --git a/scripts/run_evaluation.sh b/scripts/run_evaluation.sh
@@ -1,5 +1,9 @@
 #!/bin/bash
 
+if [ ! -d data/MedQA-USMLE ]; then
+  cd data && ./download.sh MedQA-USMLE && cd -
+fi
+
 CUDA_VISIBLE_DEVICES=0 \
     deepspeed examples/evaluation.py \
     --answer_type medmcqa \

diff --git a/scripts/run_evaluation_accelerator.sh b/scripts/run_evaluation_accelerator.sh
@@ -1,5 +1,9 @@
 #!/bin/bash
 
+if [ ! -d data/MedQA-USMLE ]; then
+  cd data && ./download.sh MedQA-USMLE && cd -
+fi
+
 CUDA_VISIBLE_DEVICES=0 accelerate launch --config_file configs/accelerator_singlegpu_config.yaml examples/evaluation.py \
     --answer_type usmle \
     --model_name_or_path gpt2-large \

diff --git a/scripts/run_evaluation_with_lora.sh b/scripts/run_evaluation_with_lora.sh
@@ -3,6 +3,11 @@
 # --model_name_or_path specifies the original huggingface model
 # --lora_model_path specifies the model difference introduced by finetuning,
 #   i.e. the one saved by ./scripts/run_finetune_with_lora.sh
+
+if [ ! -d data/alpaca ]; then
+  cd data && ./download.sh alpaca && cd -
+fi
+
 CUDA_VISIBLE_DEVICES=0 \
     deepspeed examples/evaluation.py \
     --answer_type text \

diff --git a/scripts/run_multistage_finetune.sh b/scripts/run_multistage_finetune.sh
@@ -11,6 +11,9 @@ project_dir=$(cd "$(dirname $0)"/..; pwd)
 output_dir=${project_dir}/output_models/${exp_id}
 log_dir=${project_dir}/log/${exp_id}
 dataset_path="${project_dir}/data/example_dataset/train"
+if [ ! -d ${dataset_path} ]; then
+  cd data && ./download.sh example_dataset && cd -
+fi
 
 mkdir -p ${output_dir} ${log_dir}
 

diff --git a/scripts/run_raft_align.sh b/scripts/run_raft_align.sh
@@ -11,6 +11,10 @@ project_dir=$(cd "$(dirname $0)"/..; pwd)
 output_dir=${project_dir}/output_models/${exp_id}
 log_dir=${project_dir}/log/${exp_id}
 
+if [ ! -d data/hh_rlhf ]; then
+  cd data && ./download.sh hh_rlhf && cd -
+fi
+
 mkdir -p ${output_dir} ${log_dir}
 
 export PYTHONPATH=.

diff --git a/scripts/run_reward_modeling.sh b/scripts/run_reward_modeling.sh
@@ -14,6 +14,9 @@ output_dir=${project_dir}/output_models/${exp_id}
 log_dir=${project_dir}/log/${exp_id}
 
 dataset_path=${project_dir}/data/hh_rlhf/rm/hh_rlhf_rm_training.json
+if [ ! -d data/hh_rlhf ]; then
+  cd data && ./download.sh hh_rlhf && cd -
+fi
 
 mkdir -p ${output_dir} ${log_dir}