SalsaNext is a real-time 3D LiDAR point cloud semantic segmentation model for autonomous driving, introduced by Tiago Cortinhal et al. in 2020. It combines speed, accuracy, and uncertainty estimation.

• Architecture: Builds upon SalsaNet’s encoder–decoder design, enhanced with residual dilated convolutions, pixel-shuffle upsampling, and a context module to increase receptive fields.
• Uncertainty Estimation: Incorporates Bayesian treatments to output point-wise epistemic and aleatoric uncertainties, enabling safer decision-making systems.
• Performance: Achieves state-of-the-art results on the Semantic-KITTI benchmark; accepts input size 1×5×64×2048, model size ~108 MB.
• Deployment: Provided in PyTorch and optimized for Snapdragon mobile NPUs, making it suitable for embedded and edge computing scenarios.

Source model repository: SalsaNext

When the user has fine-tuned the source model, the model conversion process must be performed again.

Users can refer to either of the following two methods to complete the model conversion:

• Using AIMO for model conversion: Click Model Conversion Reference in the Performance Reference section on the right to view the conversion steps.

• Using Qualcomm QNN for model conversion: Please refer to the Qualcomm QNN Documentation.

The model performance benchmarks and example code provided by Model Farm are all implemented based on the APLUX AidLite SDK.

For models in .bin format, you can use either of the following two inference engines to run inference on Qualcomm chips:

Inference using APLUX AidLite: please refer to the APLUX AidLite Developer Documentation

Inference using Qualcomm QNN: Please refer to the Qualcomm QNN Documentation

Inference Example Code

The inference example code is implemented using the AidLite SDK.

Click Model & Code to download the model files and the inference code package. The file structure is as follows:

/model_farm_{model_name}_aidlite
    
    |__ models # folder where model files are stored

    |__ python # aidlite python model inference example

    |__ cpp # aidlite cpp model inference example

    |__ README.md