1 post / 0 new
Join Date: 30 Jul 20
Posts: 1
Posted: Fri, 2020-07-31 21:48

I'm trying to convert a keras TensorFlow model from .pb to .dlc format using snpe-tensorflow-to-dlc.

My model is basically tensorflow.keras.applications.inception_v3.

When I run

snpe-tensorflow-to-dlc --input_network model.pb --input_dim input_1 "1,128,128,3" --out_node "final_output/Softmax" --output_path model.dlc

I get the error "Conversion failed: ERROR_TF_BATCHNORM_GLOBALNORMALIZATION_INPUT: Cannot resolve BatchNorm layer due to BatchNormWithGlobalNormalization node not having at least 4 const inputs (mean, variance, beta, scale)."

Does anyone know how to fix this? Is it because batchnorm has some trainable (non-const) parameters even after training?


(SNPE 1.40.0 and TensorFlow 1.15)

  • Up0
  • Down0

Opinions expressed in the content posted here are the personal opinions of the original authors, and do not necessarily reflect those of Qualcomm Incorporated or its subsidiaries (“Qualcomm”). The content is provided for informational purposes only and is not meant to be an endorsement or representation by Qualcomm or any other party. This site may also provide links or references to non-Qualcomm sites and resources. Qualcomm makes no representations, warranties, or other commitments whatsoever about any non-Qualcomm sites or third-party resources that may be referenced, accessible from, or linked to this site.