# rvc2

Python API: `modelconverter.platforms.rvc2`

Conversion backend for the RVC2 platform.

Holds the exporter, inferer and benchmark for RVC2. The exporter accepts an ONNX model, an OpenVINO IR pair or a TFLite model; a
TFLite model is converted to ONNX first, which the exporter itself logs as experimental. The conversion then goes through
OpenVINO: its model optimizer turns the ONNX model into an OpenVINO IR and `compile_tool` compiles that into a `.blob` -- by
default one blob per SHAVE count, packed into a single `.superblob`.

> **Note**
> These modules import the OpenVINO toolchain, which exists only inside the RVC2 Docker image, so they do not import on the host.

## Child Pages

 * [benchmark](https://docs.luxonis.com/software-v3/ai-inference/conversion/rvc-conversion/offline/modelconverter/modelconverter-api-reference/platforms/rvc2/benchmark.md)
 * [exporter](https://docs.luxonis.com/software-v3/ai-inference/conversion/rvc-conversion/offline/modelconverter/modelconverter-api-reference/platforms/rvc2/exporter.md)
 * [inferer](https://docs.luxonis.com/software-v3/ai-inference/conversion/rvc-conversion/offline/modelconverter/modelconverter-api-reference/platforms/rvc2/inferer.md)