Onnx2tf

- `pseudo_Atan`

-rtpo, --replace_to_pseudo_operators
Replace list of operators to pseudo operators.
Full name of the target operators should be given.
Currently supported operators :
Asin, Acos, Atan, Abs, PReLU, LeakyReLU, Power, GatherND, Neg, HardSwish, Erf

- [[TODO] Approximate computation implementation of Atan / Atan2 166](https://github.com/PINTO0309/onnx2tf/issues/166)
- Because of bandwidth limitations of GitHub releases, the storage from which resources are retrieved is migrated to Wasabi.

What's Changed
* `pseudo_Atan` by PINTO0309 in https://github.com/PINTO0309/onnx2tf/pull/167
* GitHub releases -> Wasabi by PINTO0309 in https://github.com/PINTO0309/onnx2tf/pull/168


**Full Changelog**: https://github.com/PINTO0309/onnx2tf/compare/1.5.40...1.5.41

There is replacement options for some operations like `Asin` and `PReLU`. However, these options are separated in argument parsing and parameter passing, which make redundant and duplicated code parts.
In this PR, those lines are simplified by using single option `replace_to_pseudo_operators`. If user give the names of target operation as list like `["acos", "hardswish"]`, each operator is replaced to pseudo operators same as before.


-rtpo, --replace_to_pseudo_operators
Replace list of operators to pseudo operators.
Full name of the target operators should be given.
Currently supported operators :
Asin, Acos, Abs, PReLU, LeakyReLU, Power, GatherND, Neg, HardSwish, Erf


What's Changed
* Simplifying operation replacement procedure to pseudo operation by Hyunseok-Kim0 in https://github.com/PINTO0309/onnx2tf/pull/165


**Full Changelog**: https://github.com/PINTO0309/onnx2tf/compare/1.5.38...1.5.40

- Fixed the number of NMS outputs to be fixed outputs in `max_output_boxes_per_class`. `-onwdt` option to generate a dynamic tensor. Taking into account that tflite's primary use is for Android, change the behavior so that it is replaced with a fixed output shape by default.
python
-onwdt, --output_nms_with_dynamic_tensor
The number of bounding boxes in the NMS output results is
not fixed at the maximum number of max_output_boxes_per_class,
but rather at the smallest possible number of dynamic tensors.
If this option is disabled, NMS output is padded to the number
set in the max_output_boxes_per_class attribute.
e.g.
disable --output_nms_with_dynamic_tensor:
output_tensor_shape: [100, 7]
enable --output_nms_with_dynamic_tensor:
output_tensor_shape: [N, 7]


- disable `-onwdt`

- enable `-onwdt`

[Yolov7-tiny to TensorflowLite comversion results in a dynamic output model incompatible with TfLite Java API 159](https://github.com/PINTO0309/onnx2tf/issues/159)

What's Changed
* To support GPU Delegate, replace all `Gather` processing part of NMS with `strided_slice` by PINTO0309 in https://github.com/PINTO0309/onnx2tf/pull/162
* Fixed the number of NMS outputs to be fixed outputs in `max_output_boxes_per_class`. `-onwdt` option to generate a dynamic tensor. by PINTO0309 in https://github.com/PINTO0309/onnx2tf/pull/164


**Full Changelog**: https://github.com/PINTO0309/onnx2tf/compare/1.5.37...1.5.38

- [BugFix] `Clip` op: Convert `tf.Tensor` min/max values to numpy arrays
- Automatic optimization of operations by replacing `Gather` and `Unsqueeze` with `Strided_Slice` when they are consecutive

What's Changed
* Automatic optimization of operations by replacing `Gather` and `Unsqueeze` with `Strided_Slice` when they are consecutive. by PINTO0309 in https://github.com/PINTO0309/onnx2tf/pull/160
* [BugFix] Clip op: Convert tf.Tensor min/max values to numpy arrays by jpowie01 in https://github.com/PINTO0309/onnx2tf/pull/161

New Contributors
* jpowie01 made their first contribution in https://github.com/PINTO0309/onnx2tf/pull/161

**Full Changelog**: https://github.com/PINTO0309/onnx2tf/compare/1.5.36...1.5.37

- Added handling of OPs that cannot be inferred from the standard onnxruntime.
- `GitHub Releases` -> `Wasabi Storage`

What's Changed
* test swap 12 -> 13 by PINTO0309 in https://github.com/PINTO0309/onnx2tf/pull/157
* Added handling of OPs that cannot be inferred from the standard onnxruntime. by PINTO0309 in https://github.com/PINTO0309/onnx2tf/pull/158


**Full Changelog**: https://github.com/PINTO0309/onnx2tf/compare/1.5.35...1.5.36

- Fixed an error with EagerTensor (np.ndarray). `-cotof`
- https://github.com/PINTO0309/onnx2tf/releases/download/1.1.27/human_segmentation_pphumanseg_2021oct.onnx
- https://github.com/PINTO0309/onnx2tf/releases/download/1.1.27/replace.json

onnx2tf -i human_segmentation_pphumanseg_2021oct.onnx -prf replace.json -cotof


File "/home/xxxxx/git/onnx2tf/onnx2tf/onnx2tf.py", line 1808, in <module>
main()
File "/home/xxxxx/git/onnx2tf/onnx2tf/onnx2tf.py", line 1757, in main
model = convert(
File "/home/xxxxx/git/onnx2tf/onnx2tf/onnx2tf.py", line 1217, in convert
onnx_tf_output_pairs = {
File "/home/xxxxx/git/onnx2tf/onnx2tf/onnx2tf.py", line 1218, in <dictcomp>
(k, v['tf_node'].name): (onnx_tensor_infos[k], tf_tensor_infos[v['tf_node'].name])
File "/usr/local/lib/python3.8/dist-packages/tensorflow/python/framework/ops.py", line 446, in __getattr__
self.__getattribute__(name)
File "/usr/local/lib/python3.8/dist-packages/tensorflow/python/framework/ops.py", line 1317, in name
raise AttributeError(
AttributeError: Tensor.name is undefined when eager execution is enabled.


What's Changed
* Fixed an error with EagerTensor (np.ndarray). `-cotof` by PINTO0309 in https://github.com/PINTO0309/onnx2tf/pull/156


**Full Changelog**: https://github.com/PINTO0309/onnx2tf/compare/1.5.34...1.5.35