video_to_sequence

First you need to download "Microsoft Video Description Corpus"
Set "video_data_path" in download_videos.py accordingly.
Download Youtube videos by running "download_videos.py"
Secondly, you need to preprocess downloaded videos
Set paths in cnn_utils.py and preprocess.py
Sample & extract features by running "preprocessing.py"
Train: train() in model.py
You might need to change the paths in "Global Parameters" area according to your environment
Test: test() in model.py

Name		Name	Last commit message	Last commit date
Latest commit History 24 Commits
.gitignore		.gitignore
LICENSE		LICENSE
MSRVTT_train_val_preprocessing.py		MSRVTT_train_val_preprocessing.py
MSVD_test_preprocessing.py		MSVD_test_preprocessing.py
MSVD_train_preprocessing.py		MSVD_train_preprocessing.py
MSVD_val_preprocessing.py		MSVD_val_preprocessing.py
MVAD_test_preprocessing.py		MVAD_test_preprocessing.py
MVAD_train_preprocessing.py		MVAD_train_preprocessing.py
MVAD_val_preprocessing.py		MVAD_val_preprocessing.py
Plot_Loss.ipynb		Plot_Loss.ipynb
README.md		README.md
cnn_util.py		cnn_util.py
cnn_util.pyc		cnn_util.pyc
download_videos.py		download_videos.py
model.py		model.py
model.pyc		model.pyc
msrvtt_ft_model.py		msrvtt_ft_model.py
msrvtt_model.py		msrvtt_model.py
msrvtt_sa_ft_model.py		msrvtt_sa_ft_model.py
msvd_ft_ls_model.py		msvd_ft_ls_model.py
msvd_ls_model.py		msvd_ls_model.py
plane.jpg		plane.jpg
preprocessing.py		preprocessing.py
test_ls_model_for_evaluation.py		test_ls_model_for_evaluation.py
test_model_for_evaluation.py		test_model_for_evaluation.py
test_msrvtt_model_for_evaluation.py		test_msrvtt_model_for_evaluation.py
test_msrvtt_sa_model_for_evaluation.py		test_msrvtt_sa_model_for_evaluation.py
video_corpus.csv		video_corpus.csv

meteora9479/video_to_sequence