Neuron Startup Script remains the same as GPU, no changes there.
Added a new triton llm backend specific neuron script.
So in theory (Not Tested), We can run neuron with djl backend, without any changes to current gpu_startup_script. Just for running the Triton llm backend, we can use the neuron_al2_startup_triton.txt with the current fmbench post startup script.
Added a _put_folder_to_instance() which can copy over folders
By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.
Neuron Startup Script remains the same as GPU, no changes there.
Added a new triton llm backend specific neuron script.
So in theory (Not Tested), We can run neuron with djl backend, without any changes to current
gpu_startup_script. Just for running the Triton llm backend, we can use theneuron_al2_startup_triton.txtwith the current fmbench post startup script.Added a
_put_folder_to_instance()which can copy over foldersBy submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.