- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Im using the command "python3 mo_tf.py --input_meta_graph <INPUT_META_GRAPH>.meta" to convert a meta file )attached with other files) and I keep getting an error ( Exception occurred during running replacer "REPLACEMENT_ID" (<class 'extensions.middle.PartialInfer.PartialInfer'>): Stopped shape/value propagation at "training/Adam/gradients/loss/dense_4_loss/Mean_grad/add" node.).
I cant seem to resolve this error.
Kindly help.
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Dearest Kobresli, Abdallah,
I noticed there's a *.pb in the folder you provided. Is it frozen ? If the *.pb is not frozen please freeze it using Tensorflow protobuf freezing instructions in OpenVino docs. Or you can also use freeze_graph.py, a Tensorflow tool. Sometimes meta graphs and checkpoints don't work well within Model Optimizer. In this case, you can freeze your *.pb and often succeed that way.
Let me know if this helps you.
For freezing you do need to know output layer names. You should already know this if you built the model yourself. Or you can fire up Tensorboard and see what the Output layer(s) are.
Thanks and I hope it helps,
Shubha
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page