Megatron-DeepSpeed icon indicating copy to clipboard operation
Megatron-DeepSpeed copied to clipboard

How to continue pre-training Bloom?

Open ShinoharaHare opened this issue 2 years ago • 2 comments

Hi I'm trying to continue pre-training the bloom-560m on my own dataset on a single GPU. I modified this script to fit my case. However, i cannot figure out how to load the checkpoint.

Is there any guide for what i'm doing?

ShinoharaHare avatar Feb 26 '23 11:02 ShinoharaHare

Hi, did you find any solution on this?

lwmlyy avatar Mar 22 '23 02:03 lwmlyy

@ShinoharaHare Hi, have you solved this problem?

noob-ctrl avatar Apr 28 '23 03:04 noob-ctrl