Open harishvs opened 4 months ago
Thanks @harishvs, I have passed the request to the Nemo maintainers and they are looking into the request.
Hello @harishvs , the issue should be resolved now with the latest 2.19 release. Please give it a try with the latest neuronx-nemo-megatron and let us know if it works for you. Thanks!
By default the llama2 example in nemo/examples/nlp/language_modeling/test_llama.sh does not resume from the last checkpoint.
Can we make this resume by default since that is a good user experience