Sequential models in DeepXDE #1647
Unanswered
dang-trinh156
asked this question in
Q&A
Replies: 1 comment
-
No. Then you need to implement the network by yourself. |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Dear Dr.Lulu and developer team and users,
I've scanned through the FAQ and search on this discussion session. All what I can find is this post:#1488.
Also, I've read the source code in Geometry, icbc, PDE, and they're all implemented on 2D-arrays (i.e. [batch_size or n_data, spatial-temporal features]). Here's an example:
The error function returns an array/tensor in 2D. Therefore, only FCNN architectures are used.
Question:
Have there been any examples, or snippets of code about the implementation of 3D inputs (i.e. [batch_size, sequence length, features]) for sequential models?
Thank you in advance!
Beta Was this translation helpful? Give feedback.
All reactions