update gpt3 125M script to be better. should exactly reproduce gpt-3 … #39
ci_gpu.yml
on: push
build-and-test-gpu
7m 46s
unit-tests-gpu
3s