xla
Add options in benchmark script to keep model on CUDA
#7193

Merged

Add options in benchmark script to keep model on CUDA #7193

vanbasten23 merged 8 commits into master from xiowei/benchmark_not_to_initialize_model_on_XLA

add how to run benchmark script for a single config without spawning …

b40411e4

Add options in benchmark script to keep model on CUDA.

797e9f3c

fix linter

edafc660

vanbasten23 requested a review from

ysiraichi 1 year ago

vanbasten23 requested a review from

zpcore 1 year ago

remove unwanted comments

009b83f9

zpcore commented on 2024-06-04

ysiraichi commented on 2024-06-05

fix comments and tests

c55e6f54

rename xla_take_cuda_model_and_data to keep_model_data_on_cuda

bbdaa9bd

vanbasten23 requested a review from

zpcore 1 year ago

fix a test

ffc9c6d6

fix the result_analyzer

3f347df2

zpcore approved these changes on 2024-06-05

vanbasten23 merged f95454ac into master 1 year ago

Reviewers

zpcore

ysiraichi

Assignees

No one assigned

Labels

None yet

Milestone

No milestone