-
In order to e.g. use neural odes, we need neural networks to operate within the dynamics. These are typically `pytorch` modules involving impure forward passes (#13), and torch functions (like an acti…
azane updated
11 months ago
-
@jstac @thomassargent30 @ChrisRackauckas
Starting with the python version as a reference: https://python.quantecon.org/sir_model.html
My thought is that there could be two lectures here:
- `M…
-
JAX is unable to obtain the Jacobian of a linear function in an efficient manner.
```python
import jax
import jax.numpy as jnp
import jax.random as jrandom
import timeit
A = jrandom.normal(j…
-
Trying to autodiff (Zygote.jl) through a neural SDE (StochasticDiffEq.jl, Lux.jl) with a `ContinuousCallback` triggering a terminate! leads to the following error
```
julia> test()
ERROR: Event was…
-
On a new install of Julia, my previously working neural SDE code no longer allows extra variables to be passed from the loss function to the callback function. The documentation says that this should …
-
With this issue I would like to get the discussion about a PDE/PDAE interface for the DifferentialEquations.jl ecosystem started.
I should also note that I might be quite biased, because I am also…
-
Hi there, I'm having OOM errors when running the summarization example on a 80GB A100 (CUDA 11.8).
I'm also getting some Tensorflow/TensorRT warnings, I'm wondering if it's related to that
```
20…
-
## Keyword: sgd
### Doubly Stochastic Models: Learning with Unbiased Label Noises and Inference Stability
- **Authors:** Authors: Haoyi Xiong, Xuhong Li, Boyang Yu, Zhanxing Zhu, Dongrui Wu, Dejin…
-
For simple cases we can just differentiate through the DDE solver but if the DDE system contains parameter-dependent C1-discontinuities the forward sensitivities have jump discontinuities which, e.g.,…
-
Hi, I'm new to this software, so this question might seem obvious to some. I'm trying to debug the Terms of Neural ODE/SDE/CDE in Diffrax with jax.debug.breakpoint(). For example, we have the code of …