Name		Name	Last commit message	Last commit date
parent directory ..
misc		misc
README.md		README.md
attention.py		attention.py
auto_encoder.py		auto_encoder.py
base_model.py		base_model.py
bc.py		bc.py
classifier.py		classifier.py
counting.py		counting.py
fc.py		fc.py
language_model.py		language_model.py
main.py		main.py
meters.py		meters.py
run_clipOrg.sh		run_clipOrg.sh
run_pubmedclip.sh		run_pubmedclip.sh
simple_cnn.py		simple_cnn.py
test.py		test.py
train.py		train.py
trainer.py		trainer.py
utils.py		utils.py

README.md

MEVF with PubMedCLIP

Follow the first and second step in README in QCR_PUBMEDCLIP.
Move data to data_RAD and data_SLAKE, respectively.
Set the CLIP_PATH to the path of fine-tuned PubMedCLIP (with respect to your encoder of choice, i.e. RN50, RN50x4, ViT)
Run the experiments for original CLIP and PubMedCLIP via run_clipOrg.sh and run_pubmedclip.sh, respectively.