Release pre-trained model
How to use
Setup
-
Download
release.zip
attached below. We will assume it will be stored under~/Downloads
directory -
Clone our repository and store
~/Downloads/release.zip
to underresult
directory. See the following instruction for more detail:
$ cd path/to/your/workspace
$ git clone https://github.com/Idein/chainer-hand-pose.git
$ cd chainer-hand-pose/result
$ cp ~/Downloads/release.zip . && unzip release.zip
$ pwd
path/to/your/workspace/chainer-hand-pose/result
$ tree
.
├── release
│ ├── detector
│ │ ├── accuracy.png
│ │ ├── best_snapshot
│ │ ├── bestmodel.npz
│ │ ├── config.ini
│ │ ├── log
│ │ └── loss.png
│ └── pose
│ ├── best_snapshot
│ ├── bestmodel.npz
│ ├── config.ini
│ ├── log
│ ├── loss.png
│ ├── plot_PCK_ppn2D.png
│ └── plot_PCK_ppn3D.png
└── release.zip
Quick Start
- We've already provided demo script inside of our repository.
- This is also an example of using actfw which is a Python API for developing Actcast apps.
$ cd path/to/your/workspace/
$ ls
chainer-hand-pose
$ cd chainer-hand-pose/src/demo
$ pip3 install actfw
$ pip3 install -e ../../
$ python3 demo.py