NIMs should be built multiplatform

Accelerated Computing DGX Spark / GB10 User Forum DGX Spark / GB10
1

Not an issue specifically regarding DGX Spark, but perhaps a more general issue regarding NIM ecosystem - NIMs should be built multi-platform (meaning for both amd64 and arm64)

I have tried first random NIM from the catalogue, stable-diffusion-3.5-large Model by Stability AI | NVIDIA NIM and it is built only for amd64, so it cannot be run on DGX Spark.

2

We are in the process of updating our NIMs models to be compatible with DGX Spark.
In the meantime, please check out our playbook on deploying NIMs on DGX Spark: Try NVIDIA NIM APIs