Jupyter notebook to train a Detectnet_v2 model with Rendered.ai user generated datasets.
TAO with Synthetic Data - Fast Demo
The ana_tao_detectnet_v2 resource contains a Jupyter Notebook and supporting configuration files to quickly train the DetectNet_v2 TAO model using Rendered.ai synthetic imagery data for a live demo. A user applies transfer learning to a Resnet-18 backbone with a targeted dataset to achieve a working object detection model in under 45 minutes. The tutorial demonstrates the effectiveness of sythetic data to train CV models.
Register now to get started with synthetic data!
More on DetectNet_v2.
Using the ANA_TAO_DETECTNET_V2 Resource
The ana_tao_detectnet_v2 resource can be deployed to Verex AI. The managed notebook environement will contain the experiment notebook file, which, when run, will create and populate a datasets folder with a prebuilt dataset from user accounts that are set up with a given content code.
NOTE: Before running any cells, make sure to select the custom kernel from ngc.
The first few cells install the Rendered.ai SDK, 'anatools.
The notebook concludes with a visulaization of performance of the trained mode on various images including real.
 |
|---|
| Effect of number of objects in a Rendered.ai graph on OD performance |
About Rendered.ai
Rendered.ai is a platform-as-a-service for synthetic data that enables data scientists to overcome the costs and challenges of acquiring and using real data for training machine learning and artificial intelligence systems.
 |
|---|
| View of Rendered.ai GUI |