Tensorflow 2.16 / Keras 3 support

[marqueurs404]

I have trained some models in Tensorflow 2.16 / Keras 3 and inference works in a normal python process. It appears to load successfully in Triton but fails upon inference with the following error logs:

Local rendezvous is aborting with status: FAILED_PRECONDITION: Could not find variable block6c_bn/gamma. This could mean that the variable has been deleted. In TF1, it can also mean the variable is uninitialized. Debug info: container=localhost, status error message=Resource localhost/block6c_bn/gamma/N10tensorflow3VarE does not exist."
[[{{function_node __inference_my_predict_6296}}{{node functional_29_1/block6c_bn_1/Cast_2/ReadVariableOp}}]]
Local rendezvous is aborting with status: FAILED_PRECONDITION: Could not find variable block6c_bn/gamma. This could mean that the variable has been deleted. In TF1, it can also mean the variable is uninitialized. Debug info: container=localhost, status error message=Resource localhost/block6c_bn/gamma/N10tensorflow3VarE does not exist."
[[{{function_node __inference_my_predict_6296}}{{node functional_29_1/block6c_bn_1/Cast_2/ReadVariableOp}}]]
[[StatefulPartitionedCall/_907]]
Local rendezvous recv item cancelled. Key hash: 2629818841554797919
Local rendezvous is aborting with status: ABORTED: Stopping remaining executors.

I noticed that the support matrix only mentioned support up to TF 2.15 for Triton, would like to know if there are any plans to support TF 2.16?

[whozwhat]

try this https://keras.io/api/models/model_saving_apis/export/