[trainer] fix: model engine vlm multi_modal_inputs to NonTensorStack #7015
wuxibin89checkpoint_converter.yml
on: pull_request
setup
6s
checkpoint_converter
1m 5s
checkpoint_converter_large_moe_models
3m 59s
cleanup
4s