Nikhil Shenoy

Fix for using Pruner trainer methods on GPU

3

### Description ### Resolves #4911 @J-shang Does this PR make sense? ### Checklist ### - [ ] test case - [ ] doc ### How to test ###

Trainer with GPU based Model fails while creating Masks

1

### Bug with GPU Model Currently, while using pruning methods like `TaylorFOWeight` Pruner, If I use a model on GPU for getting the metrics (as calculated for getting masks), it...

[Bug] RuntimeError: No backend type associated with device type cpu

22

### Bug description On upgrading torch and lightning to both 2.1.0, and running DDP leads to the following error trace, ``` # Error messages and logs here please 23 Traceback...

Fix for #411

Added vector output from skew-symmetric tensor

6

Allowing vector output from TensorNet. This will allow the use of `EquivariantScalar` or `EquivariantVectorOutput` from `TensorNet`. Resolves #297.

Gradient Checkpointing Error

1

I was trying to use [gradient checkpointing](https://pytorch.org/docs/stable/checkpoint.html) with TorchMD model. For some reason, I get this gradient mismatch whenever the warning pops up https://github.com/torchmd/torchmd-net/blob/8a1be712a397d88eed82ae12066f1533f12152ee/torchmdnet/models/utils.py#L492-L499

[Feature Request] Vector output from TensorNet

2

##### Feature Request Currently the vector output is set to `None`. This could be easily calculated based on skew symmetric matrix based on the paper. Some papers suggest using force...

Speed compared to `TorchMD-Net`

1

I wanted to train this network on the spice dataset (similar task where I want to predict forces and energy from structure). I was comparing speed of training with `torchmd-net`...

SMILES string

1

How do I access the SMILES string for any molecule?