Enable min cut partitioner by default for layout transpose optimization

google-ai-edge / ai-edge-torch

Supporting PyTorch models with the Google AI Edge TFLite runtime.

Apache License 2.0

310 stars 40 forks source link

Closed chunnienc closed 3 months ago

chunnienc commented 4 months ago

Min cut partitioner is enabled only if the graph size is smaller than a magic number determined from our model set.

Enable to better support transpose eliminations from IO transformations