Framework providing pythonic APIs, algorithms and utilities to be used with Modulus core to physics inform model training as well as higher level abstraction for domain experts
With the latest nvFuser update, one additional kernel is generated in the SiLU triple backward pass implementation. This is still much fewer than the ~35 kernels that torch generates by default, so is not a big concern. Updating the test to conform with this latest nvFuser version.
Modulus Pull Request
Description
With the latest nvFuser update, one additional kernel is generated in the SiLU triple backward pass implementation. This is still much fewer than the ~35 kernels that torch generates by default, so is not a big concern. Updating the test to conform with this latest nvFuser version.
Checklist
Dependencies
None