search

ponder-lab / Hybridize-Functions-Refactoring

Refactorings for optimizing imperative TensorFlow clients for greater efficiency.
Eclipse Public License 2.0
0 stars 0 forks source link

Potential false positive on primitive parameter #308

Closed khatchad closed 8 months ago

khatchad commented 8 months ago
subject function module relative path parameters tensor parameter primitive parameter hybrid side-effects recursive refactoring passing precondition status
samples MyConvolution2D.call tensorflow_keras_custom_layer tensorflow_keras_custom_layer/tensorflow_keras_custom_layer.py 2 TRUE TRUE FALSE FALSE FALSE CONVERT_EAGER_FUNCTION_TO_HYBRID   3

I don't see a primitive parameter here:

https://github.com/ponder-lab/samples/blob/39f7644391e664244b45c90868c804abad923eb3/tensorflow_keras_custom_layer/tensorflow_keras_custom_layer.py#L39-L57

I only see input_node here, but that should be a tensor:

https://github.com/ponder-lab/samples/blob/39f7644391e664244b45c90868c804abad923eb3/tensorflow_keras_custom_layer/tensorflow_keras_custom_layer.py#L60-L61

khatchad commented 8 months ago 
!ENTRY edu.cuny.hunter.hybridize.core 1 0 2024-01-10 11:38:22.036
!MESSAGE Found constant value: 3 for parameter of: MyConvolution2D.call().