Question
X-CUBE-AI 8.1 is out!
We are pleased to announce that the latest version of STM32Cube.AI is now available.
You can download it here.
What is new in X-CUBE-AI 8.1.0?
- TensorFlow 2.12 and Keras.io 2.12
- ONNX Runtime 1.15.0
- Host PC system: adding support for MacOS ARM system
- Include new version of the TensorFlow Lite for Microcontrollers library (SHA-1: 6a1803)
Improvements/enhancements
- Improved inference time for quantized convolutional kernels
- UI: add specific checkbox allowing to set the “–no-onnx-io-transpose” option
- CLI: enhance global model format detection, improve the report of the number of operation per layer
- Core - Code generator
- optimization of the quantized kernel implementations: CONV2D (151774), ADD (153931)
- add support for new layers:
- Einsum (bz121864), BitShift / ONNX Operators
- Support to the gather with second input variable (TFLITE & ONNX Operators)
- Support to LSTM with constant inputs not put in initializers (141640)
Major bug fixes
- (129411) TFLITE - support for transpose of batch dimension
- (152744) ONNX - enhance import of concat operator managing arbitrary inputs order
- (148917) ONNX - enhance import of resize operator considering the constant inputs
- (135654) TFLITE - enhance Dilated convolution 1D/2D (SpaceToBatchNd/BatchToSpaceNd)
- (141553) ONNX - support transpose lop between dequantize and convolution ops
- (144706) run-time lib - extend n_macc resolution to 64b
- (149752, 150344) code gen - memory allocator, enhance multiple heap support
Feel free to learn more about our STM32Cube.AI by looking into our documentation!
Attention: X-CUBE-AI works as an extension (pack) of STM32CubeMX. It requires STM32CubeMX 5.0.1 or newer versions.