Bryan Lozano

ML Suite 1.5 is the planned release to more broadly support Tensorflow. The previous support was really more of a proof of concept.

Hey, The models that we acquired from the Deephi acquisition are not being publicly released as of yet. They are pruned versions of standard models. Can you share the reason...

Hey, Even though it’s the same network architecture, the caffe model, and tensor flow model are not the same. They were trained in different frameworks, and I’m not sure that...

Hey, you can use the ml-suite AMI now. It should help you jump past a lot of set up issues. After you start, do conda activate ml-suite

Since we acquired Deephi, we are integrating new front end quantization and compilation tools. So our Keras support is pending an update. Most likely to be addressed within a month.

Hey, For ML Suite we don’t have support for U280 yet. We are working on it now. Our software supports 2018.2 XRT APIs. We are currently developing 2019.2 related code,...

I landed here for tangential reason, but was also seeing this error when trying to do conda-build for my own custom conda package. As others have suggested, removing channel priority...

Seems like it might be Windows NTTE time. https://stackoverflow.com/questions/4869769/convert-64-bit-windows-date-time-in-python/4870249#4870249

I want this too! Looks like it would need to be added to the load_conv_bn2caffe function.

This problem seems to be related to a NVIDIA/caffe version 0.15. It looks like NVIDIA fixed it in the later releases.