rknncli 0.2.2

A command line tool for parsing and displaying RKNN model information

RKNN CLI

A command line tool for parsing and displaying RKNN model information with FlatBuffers support.

Features

• Parse both FlatBuffers binary data and JSON metadata from RKNN files
• Extract detailed model information including format, compiler, runtime
• Display input/output tensor information with layout (NCHW) and data type
• Python 3.8+ compatible

Installation

From PyPI (Recommended)

pip install rknncli

Development Installation

git clone https://github.com/your-username/rknncli.git
cd rknncli
pip install -e .

Usage

rknncli <path-to-rknn-model>

Example:

rknncli assets/base-encoder.rknn

Output Format

The tool prints comprehensive model information including FlatBuffers metadata:

Model: rknn model
Target Platform: rk3588
Format: RKNPU v2
Source: ONNX
Compiler: 2.1.0+708089d1(compiler version: 2.1.0)
Runtime: rk3588
Number of graphs: 1

Input information
--------------------------------------------------------------------------------
  ValueInfo "base-mel": type FLOAT32, shape [1, 80, 3000], layout NCHW,

Output information
--------------------------------------------------------------------------------
  ValueInfo "cross_k_0": type FLOAT16, shape [1, 1500, 512], layout NCHW,
  ValueInfo "cross_v_0": type FLOAT16, shape [1, 1500, 512], layout NCHW,

Development

Prerequisites

• Python 3.8+
• FlatBuffers compiler (for schema updates)

Setup

# Clone the repository
git clone https://github.com/your-username/rknncli.git
cd rknncli

# Install in development mode
pip install -e .

# Run tests
rknncli assets/base-encoder.rknn

Updating FlatBuffers Schema

If you need to update the RKNN schema:

cd schemas
./generate_schema.sh

Publishing to PyPI

To publish a new version:

# Build the package
python -m build

# Upload to PyPI
./scripts/upload_to_pypi.sh

Changelog

See CHANGELOG.md for version history.

License

MIT License - see LICENSE file for details.