- Added **`Power`** EdgeTPU optimization process - Optimizes the **`Power`** layer to EdgeTPU only when the **`--output_edgetpu`** flag is enabled. - Replace three or more powers with **`Multiply`**. - If the exponent part is negative or decimal, use **`Power`** without replacement.
- Support for **`Einsum`** - Add **`change_equation`** option for **`Einsum`** - https://github.com/PINTO0309/openvino2tensorflow#6-7-replace-weights-or-constant-values-in-const-op-and-add-transpose-or-reshape-or-cast-or-squeeze-or-unsqueeze-just-beforeafter-the-operation-specified-by-layer_id
1.28.3
- Support for **`LogSoftmax`** - TensorRT 8.2.2 -> TensorRT 8.2.3
1.28.1
- Add **`--non_verbose`** option. - Do not show all the weight information of each layer in the conversion log.