Skip to content

Latest commit

 

History

History

guideline_network

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
__pycache__
__pycache__
 
 
data
data
 
 
docs
docs
 
 
exps
exps
 
 
llama
llama
 
 
util
util
 
 
README.md
README.md
 
 
check_the_clip_shape.py
check_the_clip_shape.py
 
 
create_json.py
create_json.py
 
 
create_json_exp1.py
create_json_exp1.py
 
 
create_json_exp2.py
create_json_exp2.py
 
 
create_json_exp3.py
create_json_exp3.py
 
 
create_json_pretrain.py
create_json_pretrain.py
 
 
create_json_vqa.py
create_json_vqa.py
 
 
demo.py
demo.py
 
 
engine_finetune.py
engine_finetune.py
 
 
engine_pretrain.py
engine_pretrain.py
 
 
evaluate.py
evaluate.py
 
 
finetune.yaml
finetune.yaml
 
 
finetune_3ch.py
finetune_3ch.py
 
 
gradio_app.py
gradio_app.py
 
 
main_finetune.py
main_finetune.py
 
 
main_pretrain.py
main_pretrain.py
 
 
playground.py
playground.py
 
 
pretrain.yaml
pretrain.yaml
 
 
requirements.txt
requirements.txt
 
 
test_pretrain.py
test_pretrain.py
 
 
test_round2_getfeature.py
test_round2_getfeature.py
 
 
tokenizer.model
tokenizer.model
 
 
train_classify.py
train_classify.py
 
 
train_classify_dataset.py
train_classify_dataset.py
 
 

README.md

LLaMA-Adapter-V2 Multi-modal

pre-train to classify modalities

CUDA_VISIBLE_DEVICES=0 python finetune_3ch.py --data_config pretrain.yaml --batch_size 3 --epochs 20 --warmup_epochs 2 --blr 10e-4 --weight_decay 0.02 --llama_path . --output_dir prostate_pretrain --pretrained_path ckpts/7fa55208379faf2dd862565284101b0e4a2a72114d6490a95e432cf9d9b6c813_BIAS-7B.pth

fine-tuning to PI-RADS

CUDA_VISIBLE_DEVICES=0 python finetune_3ch.py --data_config finetune.yaml --batch_size 3 --epochs 100 --warmup_epochs 5 --blr 10e-4 --weight_decay 0.02 --llama_path . --output_dir prostate_pretrain --pretrained_path prostate_pretrain/checkpoint-19.pth