Paddle paddle.nn.LSTM中参数time_major不起作用

fjnneemd 于 2021-11-30 发布在 Java

关注(0)|答案(3)|浏览(269)

如题，time_major=True或者False,返回的output的shape都是[batch_size, time_steps, hidden_size]

input = paddle.tensor.full((10, 100, 128), 1.0, 'float32') lstm = LSTM(128, 256, time_major=True) out, final_s = lstm(input) out.shape Out[12]: [10, 100, 256]

Paddle

来源：https://github.com/PaddlePaddle/Paddle/issues/31392

3条答案

按热度按时间

inkz8wg91#

您好，我们已经收到了您的问题，会安排技术人员尽快解答您的问题，请耐心等待。请您再次检查是否提供了清晰的问题描述、复现代码、环境&版本、报错信息等。同时，您也可以通过查看官网API文档、常见问题、历史Issue、AI社区来寻求解答。祝您生活愉快～

Hi! We've received your issue and please be patient to get responded. We will arrange technicians to answer your questions as soon as possible. Please make sure that you have posted enough message to demo your request. You may also check out the API，FAQ，Github Issue and AI community to get the answer.Have a nice day!

赞(0）回复(0）举报 2021-11-30

z4iuyo4d2#

time_major 参数会影响输出的形状的语义，输出和输入语义保持一致。
如果你的输入是 (B, T, C), 那么应该设置 time_major=False, 输出也是 (B, T, C).
如果你的输入是 (T, B, C), 那么应该设置 time_major=True, 输出也是 (T, B, C).

模型并不真正知道什么是 batch_size, 什么是 time_steps, 它只是根据协议来决定把那个当作 batch_size, 哪个当作 time_steps.
因为你的输入就是 (10, 100, 128), 设置 time_major=True 的情况下，它会认为 batch_size 是 100, time_steps 是 10.

赞(0）回复(0）举报 2021-11-30

hwamh0ep3#

理解了，多谢！

赞(0）回复(0）举报 2021-11-30