PaddleVide Quick Start

August 4, 2022 · View on GitHub

English | 简体中文

PaddleVide Quick Start

1. Installation
- 1.1 Install PaddlePaddle
- 1.2 Install PaddleVideo Whl Package
2. Easy-to-Use
- 2.1 Use by Command Line
- 2.2 Use by Python Code
3. Arguments description
4.QA

1. Installation

1.1 Install PaddlePaddle

If you have CUDA 9 or CUDA 10 installed on your machine, please run the following command to install
```
python3.7 -m pip install paddlepaddle-gpu -i https://mirror.baidu.com/pypi/simple
```
If you have no available GPU on your machine, please run the following command to install the CPU version
```
python3.7 -m pip install paddlepaddle -i https://mirror.baidu.com/pypi/simple
```

For more software version requirements, please refer to the instructions in Installation Document for operation.

1.2 Install PaddleVideo Whl Package

option1: use pypi（recommand）

pip3.7 install ppvideo==2.3.0

option2: build and install locally

python3.7 setup.py bdist_wheel
python3.7 -m pip install dist/ppvideo-2.3.0-py3-none-any.whl

2. Easy-to-Use

2.1 Use by Command Line

Run shell command：

ppvideo --model_name='ppTSM_v2' --use_gpu=False --video_file='data/example.avi'

This command use PP-TSM_v2 model to infer data/example.avi file in CPU.
The length of the example video is about 10s. When inference, the video is first divided into 16 segments according to the time axis, then extract one frame from each segment. Finally all frames are combined and feeded into the network.

Results：

Current video file: data/example.avi
        top-1 classes: [5]
        top-1 scores: [1.]
        top-1 label names: ['archery']

As you can see, use PP-TSM_v2 trained on Kinetics-400 to predict data/example.avi video，top1 prediction class_id is 5, scores is 1.0, class name is archery.

2.2 Use by Python Code

Run python code：

from ppvideo import PaddleVideo
clas = PaddleVideo(model_name='ppTSM_v2', use_gpu=False)
video_file='data/example.avi'
clas.predict(video_file)

This code use PP-TSM_v2 model to infer data/example.avi file in CPU.

Results:

Current video file: data/example.avi
        top-1 classes: [5]
        top-1 scores: [1.]
        top-1 label names: ['archery']

As you can see, use PP-TSM_v2 trained on Kinetics-400 to predict data/example.avi video，top1 prediction class_id is 5, scores is 1.0, class name is archery.

3. Arguments description

name	type	description
model_name	str	optional, model name, `'ppTSM'` or `'ppTSM_v2'`. If None, please specify the path of your inference model by args `model_file` and `params_file`.
video_file	str	required, Video file path, supported format: single video file path, or folder containing multiple videos.
use_gpu	bool	whether to use GPU，default True。
num_seg	int	The number of segments used in the TSM model, which is also the number of frames extracted from the video. 8 for `ppTSM`, 16 for `ppTSM_v2`, default 16.
short_size	int	short size of frame, default 256.
target_size	int	target size of frame, default 224.
model_file	str	optional，inference model(`.pdmodel`)path.
params_file	str	optional, inference modle(`.pdiparams`) path.
batch_size	int	Batch size, default 1.
use_fp16	bool	whether to use float16，default False.
use_tensorrt	bool	whether to use Tensorrt, default False.
gpu_mem	int	use GPU memory, default 8000.
enable_mkldnn	bool	whether to use MKLDNN, default False.
top_k	int	top_k, default 1.
label_name_path	str	This file consists the relation of class_id and class_name. Default use `data/k400/Kinetics-400_label_list.txt` of Kinetics-400. You can replace it with your own label file.

command example1：

ppvideo --model_name='ppTSM_v2' --num_seg=16 --video_file="data/mp4" --batch_size=2  --top_k=5

Results：

Current video file: data/mp4/example3.avi
        top-5 classes: [  5 345 311 159 327]
        top-5 scores: [1.0000000e+00 1.0152016e-11 8.2871061e-14 6.7713670e-14 5.0752070e-14]
        top-5 label names: ['archery', 'sword_fighting', 'skipping_rope', 'hula_hooping', 'spray_painting']
Current video file: data/mp4/example2.avi
        top-5 classes: [  5 345 311 159 327]
        top-5 scores: [1.0000000e+00 1.0152016e-11 8.2871061e-14 6.7713670e-14 5.0752070e-14]
        top-5 label names: ['archery', 'sword_fighting', 'skipping_rope', 'hula_hooping', 'spray_painting']
Current video file: data/mp4/example.avi
        top-5 classes: [  5 345 311 159 327]
        top-5 scores: [1.0000000e+00 1.0152016e-11 8.2871061e-14 6.7713670e-14 5.0752070e-14]
        top-5 label names: ['archery', 'sword_fighting', 'skipping_rope', 'hula_hooping', 'spray_painting']
Current video file: data/mp4/example1.avi
        top-5 classes: [  5 345 311 159 327]
        top-5 scores: [1.0000000e+00 1.0152016e-11 8.2871061e-14 6.7713670e-14 5.0752070e-14]
        top-5 label names: ['archery', 'sword_fighting', 'skipping_rope', 'hula_hooping', 'spray_painting']

command example1：

ppvideo --model_name='ppTSM' --num_seg=8 --video_file="data/mp4" --batch_size=2  --top_k=5

4. QA

opecv-python Installation maybe slow, you can try:

python3.7 -m pip install opencv-python==4.2.0.32 -i https://pypi.doubanio.com/simple