NVIDIA / modulus-launch

Repo of optimized training recipes for accelerating PyTorch workflows of AI driven surrogates for physical systems
Apache License 2.0
56 stars 27 forks source link

🚀[FEA]: Increment GraphCast BatchSize #130

Closed Javier-Jimenez99 closed 10 months ago

Javier-Jimenez99 commented 10 months ago

Is this a new feature, an improvement, or a change to existing functionality?

Improvement

How would you describe the priority of this feature request

Medium

Please provide a clear description of problem you would like to solve.

It can't be trained a GraphCast model with higher batchsize than 1. It is not explained why and the functions ask for a tensor with shape [N, C, H, W].

Describe any alternatives you have considered

Allow to train higher batchsizes