-
In the documentation of [monitor every multiple steps](https://brainpy.readthedocs.io/en/latest/tutorial_simulation/monitor_per_multiple_steps.html), two methods are provided. One using `brainpy.math.…
-
After upgrading to BrainPy 2.5.0, I found that training by gradient accumulation does not work in the newest version.
We can use logistic regression as an example:
```python
import numpy as np
i…
-
I need to simulate a network with trials of long durations. To avoid out-of-memory error, I divide the trial into several segments. So the simulation code has two nested for loops: the outer loop over…
-
Hello,
I commented earlier on a different issue but couldn't reopen it, so I'm opening this one. Thank you.
https://github.com/brainpy/BrainPy/issues/599#issuecomment-1918451756
-
When training a large model, it is easy to get out-fo-memory error even when the batch size is small. In PyTorch we can overcome this issue by gradient accumulation, that is, we split the batch into m…
-
Hello!
I am new to BrainPy and find it a great simulation tool!
But when I played with it, I met some errors and couldn't find solutions in the documentation.
I want to make two variables, `var…
-
Hi, BrainPy team.
In the`loss()`, I am going to slice the net's output according to the pre-setted index that stored in 'gt' , but I met some errors.
Here is the definition of `Trainer()`:
```
…
-
Hi BrainPy team!
I'd like to use BrainPy to build my model and use particle swarm optimization (PSO, by PySwarms) to find the best parameters.
However, the memory usage increased quickly when re-i…
-
Hello,
I used NEURON simulator before. In NEURON, I can use something like `record(ina)` to record sodium current.
Is there a way to monitor the current with built-in models instead of constructin…
-
In #626 we mentioned that the rotation method in delay variables does not implement an autograd functionality. However I have tested this in training and found that the parameters can be trained norma…