python - Standalone Protocol buffer file in tensorflow and its usage? -


i have .pb file alone created using freeze_graph.py other standalone .pb file don't have files in folder. till have figured out way visualize model using tensor board , fetch tensors involved in net. .pb model created using resnet architecture.

1.how make use of file make predictions on custom data have? if have process pipeline off great help.

2.also how use .pb file in keras?

  1. well load in frozen model , define input pipeline. here's an article describes in more detail. suggest have @ tensorflow serving.

  2. don't think makes sense keras uses hdf5 store model. doesn't make sense use single .pb development not use case. if want have flexibility, should use checkpoints has graph , weights epochs.


Comments

Popular posts from this blog

python Tkinter Capturing keyboard events save as one single string -

android - InAppBilling registering BroadcastReceiver in AndroidManifest -

javascript - Z-index in d3.js -