redbrick-sdk 0.0.3

RedBrick platform python SDK!

Introduction

This is an SDK to make integrating with the RedBrick platform as easy as possible. This includes uploading and downloading data as well as making your datasets easily available for training.

Desired behaviors:

• Upload:

  • upload a folder of images to a redbrick dataset

• download:

  • download images from redbrick, save as .jpg/.png or numpy arrays
  • download labels and save to specific file formats:
    • Coco format
    • darknet training format

• training:

  • Data in vectorized format ready to be converted to

Quickstart

After creating an account on app.redbrickai.com

pip install redbrick-sdk

import redbrick
import torchvision

# Setup Redbrick Global context
client = redbrick.client(org_id="asdasd", api_key="asdasd")

# create a prelabeler instance
prelabeler = redbrick.prelabeling(project_id="project_id", stage_name="stage_name")

# show an example image
prelabeler.show_sample()

# import a model
model = torchvision.models.detection.fasterrcnn_resnet50_fpn(pretrained=True)
model.eval()

# run inference and upload predictions to redbrick

Usage

redbrick.init()

Developing

pip install -e .[dev]