Occured Error When I execute export_graph method. #1071
Description
I trained classification model based on Deverta architecture using transformers package.
And I converted onnx file with my model.
It was successful.
By the way, Error was occured When I save to tensorflow file using onnx file.
How can I convert to tf file using onnx file ?
My code is simple :
tf_rep = prepare(onnx_model)
tf_rep.export_graph(TF_PATH)Error is here
File "/home/jovyan/Segmentation/onnx-tensorflow/onnx_tf/backend_tf_module.py", line 99, in __call__ *
output_ops = self.backend._onnx_node_to_tensorflow_op(onnx_node,
File "/home/jovyan/Segmentation/onnx-tensorflow/onnx_tf/backend.py", line 347, in _onnx_node_to_tensorflow_op *
return handler.handle(node, tensor_dict=tensor_dict, strict=strict)
File "/home/jovyan/Segmentation/onnx-tensorflow/onnx_tf/handlers/handler.py", line 59, in handle *
return ver_handle(node, **kwargs)
File "/home/jovyan/Segmentation/onnx-tensorflow/onnx_tf/handlers/backend/gather_elements.py", line 71, in version_13 *
return cls._common(node, **kwargs)
File "/home/jovyan/Segmentation/onnx-tensorflow/onnx_tf/handlers/backend/gather_elements.py", line 41, in _common *
axis_perm = tf.tensor_scatter_nd_update(tf.range(tf.rank(data)),
TypeError: Expected any non-tensor type, but got a tensor instead.And I analized "gather_elements.py" file.
"tf.constant([axis, 0])" is the problem.
axis variable is "Tensor("Add_18:0", shape=(), dtype=int32)"
So, It was not complete convert to constant.
Let me know How can I this problem?...!
Python, ONNX, ONNX-TF, Tensorflow version
This section can be obtained by running get_version.py from util folder.
- Python version: 3.8.12
- ONNX version: 1.11.0
- ONNX-TF version: 1.10.0
- Tensorflow version: 2.13.0
Additional context
I added this PR When I install onnx-tf package.
Because "Squeeze" and "Unsqueeze" is not supported 13 version in the latest version.
Add any other context about the problem here.