Hi,
after successfully training a brain and bring it back into unity for the Getting Started Ball Balancing tutorial I am returned the following, if I use the .byte file native to the ml-agents package it works fine, any suggestions?
TFException: Op type not registered 'ClipByValue' in binary running on REILLYAW. Make sure the Op and Kernel are registered in the binary running in this process.
TensorFlow.TFStatus.CheckMaybeRaise (TensorFlow.TFStatus incomingStatus, System.Boolean last) (at <6ed6db22f8874deba74ffe3e566039be>:0)
TensorFlow.TFGraph.Import (TensorFlow.TFBuffer graphDef, TensorFlow.TFImportGraphDefOptions options, TensorFlow.TFStatus status) (at <6ed6db22f8874deba74ffe3e566039be>:0)
TensorFlow.TFGraph.Import (System.Byte[] buffer, TensorFlow.TFImportGraphDefOptions options, TensorFlow.TFStatus status) (at <6ed6db22f8874deba74ffe3e566039be>:0)
TensorFlow.TFGraph.Import (System.Byte[] buffer, System.String prefix, TensorFlow.TFStatus status) (at <6ed6db22f8874deba74ffe3e566039be>:0)
CoreBrainInternal.InitializeCoreBrain (Communicator communicator) (at Assets/ML-Agents/Scripts/CoreBrainInternal.cs:123)
Brain.InitializeBrain (Academy aca, Communicator communicator) (at Assets/ML-Agents/Scripts/Brain.cs:209)
Academy.InitializeEnvironment () (at Assets/ML-Agents/Scripts/Academy.cs:230)
Academy.Awake () (at Assets/ML-Agents/Scripts/Academy.cs:208)
From this website, it seems someone with the same error as you fixed his issue by downgrading the TensorFlow version to 1.4. Could you try this out?
I hope this helps.
@reillydonovan You could also try to use the new TensorFlowSharp 1.7 plugin in this PR if you are interested. https://github.com/Unity-Technologies/ml-agents/pull/746
I had the same problem. No need anymore to go all the way back to 1.4. Its enough to go to 1.7 and ensure you have the latest TensorFlowSharp (Also 1.7). TensorFlow GPU 1.7 uses Cuda 9 whereas 1.4 would require 8.
I used the gpu version:
pip uninstall tensorflow-gpu
pip install tensorflow-gpu==1.7
I reran the training and all is now working fine.
nice, thanks, I reverted back to 1.4, now I'll try to go back to 1.7 :)
Thanks for reaching out to us. Hopefully you were able to resolve your issue. We are closing this due to inactivity, but if you need additional assistance, feel free to reopen the issue.
This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.
Most helpful comment
I had the same problem. No need anymore to go all the way back to 1.4. Its enough to go to 1.7 and ensure you have the latest TensorFlowSharp (Also 1.7). TensorFlow GPU 1.7 uses Cuda 9 whereas 1.4 would require 8.
I used the gpu version:
I reran the training and all is now working fine.