[InferDataLayouts] Add fallback for layout propagation

fastmachinelearning / qonnx

QONNX: Arbitrary-Precision Quantized Neural Networks in ONNX

https://qonnx.readthedocs.io/

Apache License 2.0

124 stars 39 forks source link

[InferDataLayouts] Add fallback for layout propagation #145

Open iksnagreb opened 3 weeks ago

iksnagreb commented 3 weeks ago

If there is no tensor layout annotation for the input of a non-FINN operator, annotate with the same fallback defaults as for the FINN-ops instead of propagating empty/invalid layouts.