-
Implement automatic gradient calculation
-
Hi!
This is a follow-up on #181. The use case is to evaluate the derivatives (e.g. gradient, hessian) of some loss function $\mathcal{L}$ with respect to some variable $\theta$ at the gradient equi…
-
### Description
The code below crashes during runtime.
A second, distinct runtime crash can be generated by editing one line in the reproducer below.
### Reproduction
Compile the code below in De…
-
Thanks a lot for your work! However, it seems that `backward` after `grad` is not supported yet. Here is a minimal example:
```python
# dy/dt = a b y, t = 0...1
# y1 = y0 exp(a b)
# dy1/da = b y…
-
Hi,
I am getting the following error when compiling with cmake. Any idea about how to fix the issue? Thanks.
hanxiang@hanxiang-OMEN-30L-Desktop-GT13-1xxx:~/autodiff/.build$ cmake --build . --ta…
-
```julia
using Flux
function fn_good(t0, τ, duration)
x = 1.0
t = t0
T = t0 + duration
while t < T
x += -τ*x
t += 0.01
end
x
end
f_good(τ) = fn_…
-
Hi @oschulz!
Now that [DifferentiationInterface.jl](https://github.com/gdalle/DifferentiationInterface.jl) is getting more and more mature, would you be interested in switching this package and its d…
-
I'm getting a weird typecasting issue in Enzyme. What could be the source of it?
https://fwd.gymni.ch/vATLS0
-
* Implement forward-mode AD for temporal connections; reverse-mode AD for layer connections
* Update at each time step
* Approximate BPTT without having to store any historical gradients
-
Hi,
Is some documentation available for the autodiff branch that we could have a look at to get familiar with the new wrapper. Thanks in anticipation.