Introduction

We release the entire code (both training phase & testing phase) for finetuning I3D model on UCF101.
I3D paper:Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset. Please also refer to kinetics-i3d for models and details about I3D.

Prerequisites

Software

• Ubuntu 16.04.3 LTS
• Python 2.7
• CUDA8
• CuDNN v6
• Tensorflow 1.4.1
• Sonnet

Hardware

GTX 1080 Ti

How to run

1. Clone this repo

git clone https://github.com/USTC-Video-Understanding/I3D_Finetune

2. Download kinetics pretrained I3D models

In order to finetune I3D network on UCF101, you have to download Kinetics pretrained I3D models provided by DeepMind at here. Specifically, download the repo kinetics-i3d and put the data/checkpoints folder into data subdir of our I3D_Finetune repo:

git clone https://github.com/deepmind/kinetics-i3d
cp -r kinetics-i3d/data/checkpoints I3D_Finetune/data

3. Create list files

We use list files in data/ucf101/ subdir to make the code find RGB images and flow data saved on disk. You have to adapt the list files to make sure the list files contain the right path to your data. Specifically, for RGB data, you have to update data/ucf101/rgb.txt. Each line in in this file should be in the format:

dir_name_of_imgs_of_a_video /path/to/img_dir num_imgs label

For example, if your RGB data of UCF101 is saved in '/data/user/ucf101/rgb', and there are 13320 subdirs in this folder, each subdir contains images from a video. If in subdir v_BalanceBeam_g14_c02, there are 96 images, and the ground truth of this video is 4, then the line for this subdir is:

v_BalanceBeam_g14_c02 /data/user/ucf101/rgb/v_BalanceBeam_g14_c02 96 4

Similarly, update data/ucf101/flow.txt for flow data. Note: we use one file to include x and y part of flow data, so we use {：s} in each line to placehold x or y in the data path. For example, if your flow data are placed like this:

|---tvl1_flow
|   |---x
|   |--- y

then you can write each line in flow.txt like this:

v_Archery_g01_c06 /data4/zhouhao/dataset/ucf101/tvl1_flow/{:s}/v_Archery_g01_c06 107 2

i.e, use {:s} replace x or y in path. If you are confused, please refer our code to see data loading details.

4. Train on UCF101 on RGB data and flow data

# Finetune on split1 of RGB data of UCF101
CUDA_VISIBLE_DEVICES=0 python finetune.py ucf101 rgb 1
# Finetune on split2 of flow data of UCF101
CUDA_VISIBLE_DEVICES=0 python finetune.py ucf101 flow 2 

We share our trained models on UCF101(RGB & FLOW) in GoogleDrive and BaiduDisk (password:ddar). You can download these models and put them in model folder of this repo. In this way you can skip the train commands above and directly run test in the next step.

5. test on UCF101 on RGB data and flow data

After you have trained the model, you can run the test procedure. First, please update _DATA_ROOT and _CHECKPOINT_PATHS in test.py by setting the value to right location of your dataset and your trained model generated in the previous step, respectively.
Then you can run testing using below commands:

# run testing on the split1 of RGB data of UCF101 
CUDA_VISIBLE_DEVICES=0 python test.py ucf101 rgb 1
# run testing on the split1 of flow data of UCF101
CUDA_VISIBLE_DEVICES=0 python test.py ucf101 flow 1
# run testing both on RGB and flow data of split1 of UCF101
CUDA_VISIBLE_DEVICES=0 python test.py ucf101 mixed 1

Results

Our training results on UCF-101 Split-1 are:

Training Split	RGB	Flow	Fusion
Split1	94.7%	96.3%	97.6%

Thanks to tf.Dataset API, we can achieve training speed at 1s/batch(64 frames)!

Contact

This work is mainly done by Hao Zhou (Rhythmblue) and Hezhen Hu (AlexHu123). If you have any questions, please create an issue in this repo. We are very happy to hear from you!