Introduction

This repository is for Bottom-up and Top-down Object Inference Networks for Image Captioning

Requirements

python3 tools/create_feats.py --infeats bottom_up_tsv --outfolder ./mscoco/feature/up_down_10_100

Download the annotations into the mscoco folder. More details about data preparation can be referred to self-critical.pytorch
Download coco-caption and setup the path of __C.INFERENCE.COCO_PATH in lib/config.py

bash experiments/btonet/train.sh

CUDA_VISIBLE_DEVICES=0 python3 main_test.py --folder experiments/btonet --resume model_epoch

Thanks the contribution of self-critical.pytorch and awesome PyTorch team.

Name		Name	Last commit message	Last commit date
Latest commit History 3 Commits
blocks		blocks
datasets		datasets
evaluation		evaluation
experiments/btonet		experiments/btonet
images		images
layers		layers
lib		lib
losses		losses
lr_scheduler		lr_scheduler
models		models
optimizer		optimizer
samplers		samplers
scorer		scorer
tools		tools
README.md		README.md
main.py		main.py
main_test.py		main_test.py