MegEngine: fix some bugs

[Ysllllll]

fix some bugs of megengine C++ model (.mge) visualization:

1. show the shape of the middle tensor;
2. fix scope matching model identifier (mgv2) due to possible leading information;

please help review, thanks~

[lutzroeder]

Each pair on new line to make diff easier in the future?

[Ysllllll]

Good idea, done.

[lutzroeder]

Why is ? mapped to boolean via [ 'bool', 'boolean' ], [ '?', 'boolean' ]? Also this._dataTypeName and this._dataType should be identical so toString() uses the same name as type.dataType?

[Ysllllll]

After testing, ['?', 'boolean'] here can be deleted.
The this._dataType here is mainly used to dynamically read the numerical data of the megengine.Tensor during visualization. It seems that the reading of the numerical data of the *.Tensor is no longer customized by the .js files of various frameworks? Therefore, this._dataTypeName is used to display the data type of mge.

[lutzroeder]

Is there a way to generalize this? For example, can these types be added and supported for other formats as well by the shared rendering code?

[Ysllllll]

The mapping here only involves some special data types (such as quantizeds4asymm), but there is no mapping for common data types (such as float32). For mge models in other formats (such as .tm format), the special data types they support are also included in this correspondence. So don't generalize? I wonder if I understand what you mean correctly?

[lutzroeder]

Instead of mapping quantizeds4asymm to uint8, is quantizeds4asymm a common data type that is supported by other frameworks? What would be the right name to use for all frameworks? Should dialog.Tensor support these data types? For example, there is `int4' in other frameworks, if it is not encoded in a standard way maybe it should report the standard name and not return data for rendering?

[Ysllllll]

Data type quantizeds4asymm is defined by us for quantifying data types, so it is not a common data type supported by other frameworks. Its corresponding name in dialog.Tensor should be quint8, which is also the name used by other frameworks. It has been fixed in the code.

[Ysllllll]

Hi @lutzroeder, can this mr be merged?

[lutzroeder]

@Ysllllll can you rebase to TIP. The mgbtest0 and ./tools/megengine changes should be no longer needed. If not, please share a small repro. The data type support is still an open issue. For now, use the specific name if there is no general name to get unblocked. This will result in the uint8 not getting rendered which can be addressed in a later change.

[Ysllllll]

OK. Yes,

• for ./tools/megengine, its changes is no longer needed.
• for mgbtest0, need to modify a little bit, that is, remove the if statement ( if (size === (stream.length - position - 4))).
  This is mainly because the current model file also contains data, and the size of the record is only the size of the model, without adding the size of the data.
  Therefore, the correct expression should be size === (stream.length - position - 4) - length(inputdata), but since the data size is unknown, the if statement needs to be removed at this time.
  Here is an example:
  add_with_data.zip