OFA-OCR / run_scripts /vqa /evaluate_vqa_zeroshot.sh
JustinLin610's picture
first commit
ee21b96
#!/usr/bin/env bash
# This script evaluates pretrained OFA-Large checkpoint on zero-shot open-domain VQA task.
# The port for communication. Note that if you want to run multiple tasks on the same machine,
# you need to specify different port numbers.
export MASTER_PORT=8082
user_dir=../../ofa_module
bpe_dir=../../utils/BPE
# val or test
split=$1
data=../../dataset/vqa_data/vqa_${split}.tsv
path=../../checkpoints/ofa_large.pt
result_path=../../results/vqa_${split}_zeroshot
selected_cols=0,5,2,3,4
CUDA_VISIBLE_DEVICES=0,1,2,3 python3 -m torch.distributed.launch --nproc_per_node=4 --master_port=${MASTER_PORT} ../../evaluate.py \
${data} \
--path=${path} \
--user-dir=${user_dir} \
--task=vqa_gen \
--selected-cols=${selected_cols} \
--bpe-dir=${bpe_dir} \
--patch-image-size=480 \
--prompt-type='none' \
--batch-size=8 \
--log-format=simple --log-interval=10 \
--seed=7 \
--gen-subset=${split} \
--results-path=${result_path} \
--fp16 \
--zero-shot \
--beam=20 \
--unnormalized \
--temperature=1.0 \
--num-workers=0