🔥🔥 Kaolin-Wisp + WandB Demo 🪄🐝

Install Kaolin Core and Kaolin Wisp

Download Sample Data for a V8 Model Engine

Train VQAD

A great aspect of Kaolin Wisp is that it comes with the goodness of Weights & Biases integrated with itself!!!

To track training and validation metrics, render 3D interactive plots, reproduce your configurations and results, and many more features in your Weights & Biases workspace just add the additional flag --wandb_project <your-project-name> when initializing the training script.

The complete list of features supported by Weights & Biases:

• Log training and validation metrics in real time.

• Log system metrics in real time.

• Log RGB, RGBA, Depth renderings etc. during training.

• Log interactive 360 degree renderings post training in all levels of detail.

• Log model checkpoints as Weights & Biases artifacts.

• Sync experiment configs for reproducibility.

• Host Tensorboard instance inside Weights & Biases run.

The full list of optional arguments related to logging on Weights & Biases include:

• --wandb_project: Name of Weights & Biases project

• --wandb_run_name: Name of Weights & Biases run [Optional]

• --wandb_entity: Name of Weights & Biases entity under which your project resides [Optional]

• --wandb_viz_nerf_angles: Number of angles in the 360 degree renderings [Optional, default set to 20]

• --wandb_viz_nerf_distance: Camera distance to visualize Scene from for 360 degree renderings on Weights & Biases [Optional, default set to 3]

If you wish to train using one of the numerous scenes from the RTMV Dataset, you can replace the gdown URL with one of the tar files from here. You also need to change the --dataset-path paramter while training to the respective path of the model that you wish to train on.