Open DLuminary opened 3 years ago
Hi, this value is used to avoid to have big neural network outputs during the first iterations and thus to train a little bit faster.
A pretty good value is to take the standard deviation of the gyro error.
En 4 jul. 2021 7:28, en 7:28, DLuminary @.***> escribió:
Hello Brossar and thanks for sharing your work.
I'm trying to use your method with a different dataset. In order to train and test the network - it seems like gyro_std needs to be defined. How can one obtain the values for some arbitrary dataset?
Thanks again!
-- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/mbrossar/denoise-imu-gyro/issues/8
Thanks for your rapid reply @mbrossar
Some questions regarding the main run file - in dataset_params: 1) Does N represent approximately the window size? (in your paper it is 448 and it seems like 500 is used in the code) 2) Do min_train_freq and max_train_freq represent the loss function factors? 3) When using data with 100Hz measurements instead of 200Hz, would you advise lowering the window size and loss factors? What other changes would you suggest?
Best regards
感谢您的快速回复@mbrossar
关于主运行文件的一些问题 - 在 dataset_params 中:
- N 是否近似代表窗口大小?(在你的论文中它是 448 并且似乎在代码中使用了 500)
- min_train_freq 和 max_train_freq 代表损失函数因子吗?
- 当使用 100Hz 测量而不是 200Hz 的数据时,您会建议降低窗口大小和损耗因子吗?您还建议进行哪些其他更改?
最好的祝福
Hello, do you still have the corresponding optimization parameters and data set of this code? The author's links have all failed. Thank you!
Thanks for your rapid reply @mbrossar
Some questions regarding the main run file - in dataset_params:
- Does N represent approximately the window size? (in your paper it is 448 and it seems like 500 is used in the code)
- Do min_train_freq and max_train_freq represent the loss function factors?
- When using data with 100Hz measurements instead of 200Hz, would you advise lowering the window size and loss factors? What other changes would you suggest?
Best regards
Hello, do you still have the corresponding optimization parameters and data set of this code? The author's links have all failed. Thank you!
Hello Brossar and thanks for sharing your work.
I'm trying to use your method with a different dataset. In order to train and test the network - it seems like gyro_std needs to be defined. How can one obtain the values for some arbitrary dataset?
Thanks again!