xml/bin to tflite conversion error : Dimension size must be evenly divisible by 2 but is 17

zhenzhen1003 commented 2 years ago

Issue Type

Others

OS

Windows

OS architecture

x86_64

Programming Language

Python

Framework

OpenVINO, ONNX, TensorFlow

Download URL for ONNX / OpenVINO IR

robust.zip robust-onnx.zip

Convert Script

From openvino to tensorflow script : python E:/Desk/openvino2tensorflow-main/openvino2tensorflow/openvino2tensorflow.py --model_path D:/tflite/robust.xml --model_output_path D:/tflite --output_no_quant_float32_tflite --output_weight_quant_tflite

From onnx to openvino script: python mo.py --framework=onnx --data_type=FP32 --output_dir=D:\public\FP32 --model_name=robust --input=src --output=pha --scale_values=src[255,255,255] --reverse_input_channels --input_model=D:\public\robust.onnx --input_shape [1,3,1080,1920]

Description

I have succeeded converting model from onnx to xml, but getting error in tflite conversion. Onnx and xml/bin models are well working. Thanks for your read and hope you can solve my problem.

Relevant Log Output

====================================================================================
layer_type: Const
layer_id: 539
tf_layers_dict_shape: (2,)
tf_layers_dict_value: [64 64]
====================================================================================
ERROR: Dimension size must be evenly divisible by 2 but is 17
        Number of ways to split should evenly divide the split dimension for '{{node lambda_1/Split}} = Split[T=DT_FLOAT, num_split=2](lambda_1/Split/split_dim, Placeholder)' with input shapes: [], [1,17,30,128] and with computed input tensors: input[0] = <-3>.
ERROR: model_path  : D:/tflite/robust.xml
ERROR: weights_path: D:/tflite/robust.bin
ERROR: layer_id    : 540
ERROR: The trace log is below.
Traceback (most recent call last):
  File "E:\miniconda3\envs\openvino\lib\site-packages\tensorflow\python\framework\ops.py", line 1880, in _create_c_op
    c_op = pywrap_tf_session.TF_FinishOperation(op_desc)
tensorflow.python.framework.errors_impl.InvalidArgumentError: Dimension size must be evenly divisible by 2 but is 17
        Number of ways to split should evenly divide the split dimension for '{{node lambda_1/Split}} = Split[T=DT_FLOAT, num_split=2](lambda_1/Split/split_dim, Placeholder)' with input shapes: [], [1,17,30,128] and with computed input tensors: input[0] = <-3>.

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "E:/Desk/openvino2tensorflow-main/openvino2tensorflow/openvino2tensorflow.py", line 5120, in convert
    )(tf_layers_dict[get_tf_edges_from(tf_edges, layer_id, 0)])
  File "E:\miniconda3\envs\openvino\lib\site-packages\tensorflow\python\keras\engine\base_layer.py", line 970, in __call__
    input_list)
  File "E:\miniconda3\envs\openvino\lib\site-packages\tensorflow\python\keras\engine\base_layer.py", line 1108, in _functional_construction_call
    inputs, input_masks, args, kwargs)
  File "E:\miniconda3\envs\openvino\lib\site-packages\tensorflow\python\keras\engine\base_layer.py", line 840, in _keras_tensor_symbolic_call
    return self._infer_output_signature(inputs, args, kwargs, input_masks)
  File "E:\miniconda3\envs\openvino\lib\site-packages\tensorflow\python\keras\engine\base_layer.py", line 880, in _infer_output_signature
    outputs = call_fn(inputs, *args, **kwargs)
  File "E:\miniconda3\envs\openvino\lib\site-packages\tensorflow\python\keras\layers\core.py", line 919, in call
    result = self.function(inputs, **kwargs)
  File "E:/Desk/openvino2tensorflow-main/openvino2tensorflow/openvino2tensorflow.py", line 5112, in split_tensor
    return tf.raw_ops.Split(axis=axis, value=x, num_split=num_split)
  File "E:\miniconda3\envs\openvino\lib\site-packages\tensorflow\python\util\tf_export.py", line 404, in wrapper
    return f(**kwargs)
  File "E:\miniconda3\envs\openvino\lib\site-packages\tensorflow\python\ops\gen_array_ops.py", line 10040, in split
    "Split", split_dim=axis, value=value, num_split=num_split, name=name)
  File "E:\miniconda3\envs\openvino\lib\site-packages\tensorflow\python\framework\op_def_library.py", line 750, in _apply_op_helper
    attrs=attr_protos, op_def=op_def)
  File "E:\miniconda3\envs\openvino\lib\site-packages\tensorflow\python\framework\func_graph.py", line 601, in _create_op_internal
    compute_device)
  File "E:\miniconda3\envs\openvino\lib\site-packages\tensorflow\python\framework\ops.py", line 3565, in _create_op_internal
    op_def=op_def)
  File "E:\miniconda3\envs\openvino\lib\site-packages\tensorflow\python\framework\ops.py", line 2042, in __init__
    control_input_ops, op_def)
  File "E:\miniconda3\envs\openvino\lib\site-packages\tensorflow\python\framework\ops.py", line 1883, in _create_c_op
    raise ValueError(str(e))
ValueError: Dimension size must be evenly divisible by 2 but is 17
        Number of ways to split should evenly divide the split dimension for '{{node lambda_1/Split}} = Split[T=DT_FLOAT, num_split=2](lambda_1/Split/split_dim, Placeholder)' with input shapes: [], [1,17,30,128] and with computed input tensors: input[0] = <-3>.
ERROR: Please refer to 6-7 in the README first. https://github.com/PINTO0309/openvino2tensorflow

Source code for simple inference testing code

python E:/Desk/openvino2tensorflow-main/openvino2tensorflow/openvino2tensorflow.py --model_path D:/tflite/robust.xml --model_output_path D:/tflite --output_no_quant_float32_tflite --output_weight_quant_tflite

PINTO0309 commented 2 years ago

See.

ERROR: Please refer to 6-7 in the README first. https://github.com/PINTO0309/openvino2tensorflow

https://github.com/PINTO0309/PINTO_model_zoo/tree/main/242_RobustVideoMatting

zhenzhen1003 commented 2 years ago

See.
ERROR: Please refer to 6-7 in the README first. https://github.com/PINTO0309/openvino2tensorflow
https://github.com/PINTO0309/PINTO_model_zoo/tree/main/242_RobustVideoMatting

Thank you very much for replying to me in such a short time. I'm sorry I didn't notice this existing transformation. I'm sorry to trouble you.

PINTO0309 / openvino2tensorflow