Hello, I am an employee (Removed by moderator). We have been working on a VWW (Visual Wake Words) model and have developed a working model for the Brainchip Akida and a TFLite model for the Cortex-M4. We want to test this TFLite model on an STM32F746G Disco board, which has a Cortex M7 processor. However, while running the application, the target board halts at the allocate_tensor function. To troubleshoot this issue, I added print logs inside the allocate_tensors function and found that the issue is at the function call graph_.PrepareSubgraphs(). I am not sure what the issue is and would appreciate any help in resolving it.
Welcome to the Tensorflow Forum!
allocate_tensor possibly fail because of memory size on the board or
in-correct input shape during the inference. Could you please share standalone to reproduce the issue?