[Typing][B-94] Add type annotations for python/paddle/hapi/dynamic_flops.py

[enkilee]

PR Category

User Experience

PR Types

Improvements

Description

• [Type Hints] 为公开 API 标注类型提示信息 #65008
  B94
  @megemini

[paddle-bot]

你的PR提交成功，感谢你对开源项目的贡献!
请关注后续CI自动化测试结果，详情请参考Paddle-CI手册。
Your PR has been submitted. Thanks for your contribution!
Please wait for the result of CI firstly. See Paddle CI Manual for details.

[megemini]

另外，改完之后要看一下 CI 的示例代码运行结果 ～

目前示例代码输出结果不一致的情况，有可能是 numpy 版本不一致导致的，可以考虑 SKIP 掉 table 的打印 ～ 原示例：

            >>> FLOPs = paddle.flops(lenet,
            ...                      [1, 1, 28, 28],
            ...                      custom_ops= {nn.LeakyReLU: count_leaky_relu},
            ...                      print_detail=True)
            >>> print(FLOPs)
            <class 'paddle.nn.layer.conv.Conv2D'>'s flops has been counted
            <class 'paddle.nn.layer.activation.ReLU'>'s flops has been counted
            Cannot find suitable count function for <class 'paddle.nn.layer.pooling.MaxPool2D'>. Treat it as zero FLOPs.
            <class 'paddle.nn.layer.common.Linear'>'s flops has been counted
            +--------------+-----------------+-----------------+--------+--------+
            |  Layer Name  |   Input Shape   |   Output Shape  | Params | Flops  |
            +--------------+-----------------+-----------------+--------+--------+
            |   conv2d_0   |  [1, 1, 28, 28] |  [1, 6, 28, 28] |   60   | 47040  |
            |   re_lu_0    |  [1, 6, 28, 28] |  [1, 6, 28, 28] |   0    |   0    |
            | max_pool2d_0 |  [1, 6, 28, 28] |  [1, 6, 14, 14] |   0    |   0    |
            |   conv2d_1   |  [1, 6, 14, 14] | [1, 16, 10, 10] |  2416  | 241600 |
            |   re_lu_1    | [1, 16, 10, 10] | [1, 16, 10, 10] |   0    |   0    |
            | max_pool2d_1 | [1, 16, 10, 10] |  [1, 16, 5, 5]  |   0    |   0    |
            |   linear_0   |     [1, 400]    |     [1, 120]    | 48120  | 48000  |
            |   linear_1   |     [1, 120]    |     [1, 84]     | 10164  | 10080  |
            |   linear_2   |     [1, 84]     |     [1, 10]     |  850   |  840   |
            +--------------+-----------------+-----------------+--------+--------+
            Total Flops: 347560     Total Params: 61610
            347560

改为

            >>> FLOPs = paddle.flops(lenet,
            ...                      [1, 1, 28, 28],
            ...                      custom_ops= {nn.LeakyReLU: count_leaky_relu},
            ...                      print_detail=True)
            >>> # doctest: +SKIP('numpy print with different version')
            <class 'paddle.nn.layer.conv.Conv2D'>'s flops has been counted
            <class 'paddle.nn.layer.activation.ReLU'>'s flops has been counted
            Cannot find suitable count function for <class 'paddle.nn.layer.pooling.MaxPool2D'>. Treat it as zero FLOPs.
            <class 'paddle.nn.layer.common.Linear'>'s flops has been counted
            +--------------+-----------------+-----------------+--------+--------+
            |  Layer Name  |   Input Shape   |   Output Shape  | Params | Flops  |
            +--------------+-----------------+-----------------+--------+--------+
            |   conv2d_0   |  [1, 1, 28, 28] |  [1, 6, 28, 28] |   60   | 47040  |
            |   re_lu_0    |  [1, 6, 28, 28] |  [1, 6, 28, 28] |   0    |   0    |
            | max_pool2d_0 |  [1, 6, 28, 28] |  [1, 6, 14, 14] |   0    |   0    |
            |   conv2d_1   |  [1, 6, 14, 14] | [1, 16, 10, 10] |  2416  | 241600 |
            |   re_lu_1    | [1, 16, 10, 10] | [1, 16, 10, 10] |   0    |   0    |
            | max_pool2d_1 | [1, 16, 10, 10] |  [1, 16, 5, 5]  |   0    |   0    |
            |   linear_0   |     [1, 400]    |     [1, 120]    | 48120  | 48000  |
            |   linear_1   |     [1, 120]    |     [1, 84]     | 10164  | 10080  |
            |   linear_2   |     [1, 84]     |     [1, 10]     |  850   |  840   |
            +--------------+-----------------+-----------------+--------+--------+
            Total Flops: 347560     Total Params: 61610
            >>> # doctest: -SKIP
            >>> print(FLOPs)
            347560

也就是说，只检查 FLOPs 的数值正确性 ～

另外，typing 有可能有需要 ignore 的地方，等 CI 结果再说吧～

[megemini]

LGTM ～

[SigureMo]

LGTMeow