[ONNXRuntimeError] : 9 : NOT_IMPLEMENTED : Could not find an implementation for the node BatchNormalization(6)

[dashesy]

Describe the bug
Try an inference session with older model

sess = rt.InferenceSession("old_tag.onnx")

RuntimeError: [ONNXRuntimeError] : 9 : NOT_IMPLEMENTED : Could not find an implementation for the node BatchNormalization(6)

System information

• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Ubuntu16
• ONNX Runtime installed from (source or binary): pip
• ONNX Runtime version: 0.4.0
• Python version: 3.6

To Reproduce
Describe steps/code to reproduce the behavior:

Expected behavior
Backward compatibility, or update changelog tot mention breaking change

Additional context
Add any other context about the problem here.

[hariharans29]

Hi Ehsan Azar (@dashesy),

Is your model an opset 6 conformant model ? It seems to be looking for a BatchNormalization node with opset version 6. Here, under the section "Backwards compatibility", it says "All versions of ONNX Runtime will support ONNX opsets all the way back to (and including) opset version 7". So, I think this is expected.

[hariharans29]

Closing as it seems this is by design since opset 6 is not supported per the versioning doc. Please feel free to re-open if you feels this was incorrectly interpreted.