# 0.5.0

## Release Information

- OS support: Tested with Windows-x86 (10), Windows-arm64 (10), Mac-arm64 (14), Linux (Ubuntu 22.04)
- QAIRT SDK: Tested with versions (2.36.0, 2.37.0)
- Model Visualization format support: ONNX, TensorFlow, PyTorch (1.x, 2.x), TFLite, Executorch Program, DLC

## Upgrade from < 0.4.0 to 0.5.0 Notes

- The CLI for QAIRT Visualizer has changed from <cite>qairt_visualizer</cite> to <cite>qairt-visualizer</cite> to better align
with the package naming and align with naming conventions for CLI tools

## Highlights

- Node grouping: adds support to organize nodes into expandable layers within the graph view

    - Nodes are grouped either with namespaces defined in models and/or using synthetic layer grouping
    - Node namespace grouping is supported for TensorFlow, TFLite, and ONNX models. Synthetic layer grouping
is supported for all models.
    - Note: This replaces the existing subgraph view for large models in QAIRT Visualizer. Any sized model
can be visualized with node grouping without the need to use the subgraph view.
- DLC quantization encodings support: adds support to visualize per node quantization encodings data for DLCs
- QNN HTP Analysis Summary (QHAS) UI enhancements

    - Support for multi-core HTP QHAS analytics
    - Added sorting options for applicable charts
    - Added restore and axis label toggle buttons to all charts

## Known Issues

- If two or more tensor outputs of a multi-output node go into the same consumer node, the search functionality can fail to highlight the correct node/edge connections

    - Recommendation for such nodes is to search and select the producing node, not its output tensors

Last Published: Aug 19, 2026

[Previous Topic
Highlights](https://docs.qualcomm.com/bundle/publicresource/80-87189-1/topics/0_4_1.md) [Next Topic
0.5.1](https://docs.qualcomm.com/bundle/publicresource/80-87189-1/topics/0_5_1.md)