modulus
modulus copied to clipboard
NVIDIA
Open-source deep-learning framework for building, training, and fine-tuning deep learning models using state-of-the-art Physics-ML methods
### Version 0.6.0 ### On which installation method(s) does this occur? Docker ### Describe the issue `NCCL_ASYNC_ERROR_HANDLING` environment variable (set [here](https://github.com/NVIDIA/modulus/blob/7a631ac01f80ba7bb70d8dee49c92ac234e624e9/modulus/distributed/manager.py#L335)) has been deprecated in favor of `TORCH_NCCL_ASYNC_ERROR_HANDLING` https://github.com/pytorch/pytorch/pull/114077 It...
### Version 0.6.0 ### On which installation method(s) does this occur? Docker, Pip ### Describe the issue The inference of CorrDiff, using Nvidia Inference Package, i.e., a pretrained model and...
### Version 24.04 ### On which installation method(s) does this occur? Docker ### Describe the issue Converted docker container to apptainer. When I ran training script, I got "too many...
### Version 0.7.0a ### On which installation method(s) does this occur? Docker, Source ### Describe the issue **Analysis**: The error comes because `model_kwargs` are double nested when passed to the...
### Version 0.7.0a ### On which installation method(s) does this occur? Docker, Source ### Describe the issue **Analysis**: Dimensions of model and/or data mismatch. This only affects `edmv2` ### Minimum...
### Version 0.7.0a ### On which installation method(s) does this occur? Docker, Source ### Describe the issue **Analysis**: Line `training_loop.py:191` calls `modulus.metrics.diffusion.EDMLossSR` ```python loss_fn = construct_class_by_name(**loss_kwargs, **interface_kwargs) ``` with `loss_kwargs...
### Version 0.7.0a ### On which installation method(s) does this occur? Docker, Source ### Describe the issue **Analysis**: `modulus/modulus/metrics/diffusion/__init__.py` does not importe the `EDMLossSR` is that intentional. **Proposed fix**: Add...
### Version 0.6.0 ### On which installation method(s) does this occur? Docker, Source ### Describe the issue Error due to args mismatch between EDMPrecondSRV2 and CorrDiff training ### Minimum reproducible...
