torchdyn
torchdyn copied to clipboard
DiffEqML
A PyTorch library entirely dedicated to neural differential equations, implicit models and related numerical methods
Adds discrete cosine transforms (type-I and II for now) and an `nn.Module` API. `Explicit` transforms precompute the matrix for efficiency at a memory cost. Starting point: https://github.com/zh217/torch-dct/blob/master/torch_dct/_dct.py. This is prep...
Several methods and classes in `torchdyn` would benefit from `vjp/jvp`: `LNN`, `MSForward` and others. We should decide on a way to integrate the functional API of `PyTorch` and then merge...
I am noticing that the training just stalls/stops at a certain epoch for some reason. No errors/explanation. Screenshot below. It got stuck at this epoch and has been here for...
I'm not sure if this is a bug, an error in my code, or by design, but my ODE errors out unless `input_dim == output_dim` at all steps within the...
Hi, thanks for the great library. I am interested in the ODE of the form dz(s)/ds = f_theta (s, r, z(s)) where r can be a vector independent of input...
Hi, First of all thank you for this nice library. I really enjoyed using it. I was wondering how can I figure out the number of steps taken by an...
Hi, I am currently working with the `torchdyn` package and I am getting an error that I cannot really explain: ``` File "/home/maxh/miniconda3/envs/deepqmc/lib/python3.8/site-packages/torch/autograd/function.py", line 87, in apply return self._forward_cls.backward(self, *args)...
Tutorial for this [paper](https://arxiv.org/abs/2106.04165). `odeint_hybrid` is already in. @frankschae On our end, it would be good to extend the NHA tutorial with adjoint sensitivity through the events. ping since you're...