You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Since most "Compute" deployments will be DC bound and headless, you need to incorporate a CLI version of Morpheus that does not require a GUI
All the GPU resources should be for the AI workload and not consumed by any GUI, most of our platforms go into datacenters where the only allowed access is SSH. Installing a GUI also screws with NVIDIA drivers and CUDA.
I'd also like to ask about other GPU Vendors, such as the Intel GPU MAX which uses C++, and also the new AI Accelerators on the 4th and 5th Gen XEONs (Advanced Matrix Extensions), which are much more efficient at inferencing than the GPU
The text was updated successfully, but these errors were encountered:
I'm interested in working on this. I strongly support the fact that there shouldn't be a need to setup an electron application in order to use the project, as electron comes with its benefits and tradeoffs.
How many hours of work to do you think this would take @YashKumarVerma ?
Seems worth doing, though maybe closer to when the Compute phase 2 is gearing up to go live in the coming months.
Also be sure to see the latest 0.0.6 version in Beta.
Since most "Compute" deployments will be DC bound and headless, you need to incorporate a CLI version of Morpheus that does not require a GUI
All the GPU resources should be for the AI workload and not consumed by any GUI, most of our platforms go into datacenters where the only allowed access is SSH. Installing a GUI also screws with NVIDIA drivers and CUDA.
I'd also like to ask about other GPU Vendors, such as the Intel GPU MAX which uses C++, and also the new AI Accelerators on the 4th and 5th Gen XEONs (Advanced Matrix Extensions), which are much more efficient at inferencing than the GPU
The text was updated successfully, but these errors were encountered: