https://github.com/zetane/coco_viz

https://github.com/zetane/coco_viz

Last synced: 36 minutes ago
JSON representation

• Host: GitHub
• URL: https://github.com/zetane/coco_viz
• Owner: zetane
• License: mit
• Created: 2021-02-23T21:05:02.000Z (almost 4 years ago)
• Default Branch: main
• Last Pushed: 2021-03-03T16:45:43.000Z (almost 4 years ago)
• Last Synced: 2024-11-10T14:08:33.910Z (about 2 months ago)
• Language: Python
• Size: 40.4 MB
• Stars: 3
• Watchers: 4
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# coco_viz

## Requirements

* `git`

* `git-lfs` [extension to Git](https://git-lfs.github.com/).

* Zetane [Viewer](https://zetane.com/gallery)

* Python 3

* Install the module requirements from `requirements.txt` (PyPi listing). This includes the Zetane Python module.

## Setup

The repository refers to Coco and datasets. They need a one-time initialization:

```

git submodule init

git lfs pull

pip install -r requirements.txt

```

The Git submodule integrates the Coco API, and LFS gets the datasets.

## Run

First, please start the Zetane viewer. This should open the viewer on a blank universe. Then:

```

python test_torch.py

```

The script will use the datasets to train a model and generate visualizations in the Zetane Viewer.

## Legacy Documentation

Run the sample

```git submodule init # fetches cocoapi

sh extract.sh      # extracts small subset of validation data

cd cocoapi/PythonAPI

python setup.py build_ext install # installs the cocoapi package in your venv or conda env

```

Run Zetane Engine via command line:

```

open /Applications/Zetane.app --args --server

```

Run sample script

```

python torch_test.py

```