pyarmnn: reshape input size TFlite and ONNX

ARM-software / armnn

Arm NN ML Software. The code here is a read-only mirror of https://review.mlplatform.org/admin/repos/ml/armnn

https://developer.arm.com/products/processors/machine-learning/arm-nn

MIT License

1.15k stars 307 forks source link

pyarmnn: reshape input size TFlite and ONNX #584

Closed helion-du-mas-des-bourboux-thales closed 2 years ago

helion-du-mas-des-bourboux-thales commented 2 years ago

Is there a way in pyArmNN TFLite or ONNX parser to change the input and output tensors. For example in TensorFlow, one would run the folowing with a dynamicly defined model

model.resize_tensor_input(input_details[0]['index'], x.shape)
model.resize_tensor_input(output_details[0]['index'], y.shape)

input_details = model.get_input_details()
output_details = model.get_output_details()

model.allocate_tensors()

Colm-in-Arm commented 2 years ago

Hi @helion-du-mas-des-bourboux-thales,

Sorry but Arm NN does not currently support dynamic size for input or output tensors.

Colm.