-
Trying to figure out a converter between e3nn and GElib. Just putting this here while I try to figure out the normalization difference
```python
import torch
import gelib
torch.manual_seed(0)
f…
-
This is a feature request to ultimately remove the version pin of `e3nn==0.4.4` in `setup.cfg` to allow for more flexible versions (if possible).
-
The error thrown: AttributeError: jax.core.pp_eqn_rules was removed in JAX v0.4.34. Which makes sense in the newest release: https://github.com/jax-ml/jax/releases. Which specifies the removal of the …
-
@atharvchagi This might be a good first issue to get your feet wet. If you look at e3nn/e3nn-jax, there's a lot of complains about not supporting lmax>=10 spherical_harmonics. I think the machinery sh…
-
I am trying to recreate my nequip-torch model to nequip-jax. While the number of parameters matches exactly when I create the model, the numerical values do not. SO I am trying to assemble the layers …
-
I have this [reproducer](https://github.com/TaufeqRazakh/graphnorm) repository to demonstrate the graph breaks with pt2 in norm function in [e3nn](https://github.com/e3nn/e3nn/blob/main/e3nn/o3/_norm.…
-
### 🐛 Describe the bug
```python
import torch
torch._dynamo.config.capture_scalar_outputs = True
def my_arithmetic(a, b):
wrk = torch.zeros(a.size(0))
for i in range(a.size(0)):
…
-
### 🐛 Describe the bug
When you have a function that uses torch.sparse, torch.compile can compile it, but not with "fullgraph=True"
### Error logs
(e3nn_env) aglover@login36:~/e3nn_benchmarking> p…
-
Hi,
Thanks for your work. May I know the e3nn version you used in gert? Or would you upload the relevant code into the empty e3nn directory? I noticed that the scripts imported this package via i…
-
Recently, the `flax` library has unveiled its latest neural network programming framework, [`nnx`](https://flax.readthedocs.io/en/latest/nnx/index.html). Compared to the original `linen`, `nnx` boasts…