Many thanks for the quick reply! The RNN seems like the relevant example. It seems that to get what I want, I need a function that has all output variables as inputs also, even if the function doesn’t act on some of them? I think I can live with that.
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| Using PyTensor.scan to loop over multidimensional array | 13 | 1365 | June 1, 2023 | |
|
Constructing a matrix within each step of `pytensor.scan`
|
4 | 49 | January 28, 2025 | |
| Nested scan outer loop only iterates over first step | 4 | 201 | February 16, 2024 | |
| How to use pytensor's scan function | 5 | 1531 | February 27, 2023 | |
| Nan when using two outputs in pytensor.scan | 2 | 208 | December 5, 2023 |