XinliYu

This is to test the outputs are the same before and after simplification. The two "print" lines in the following code do the checking. ``` def load_data(dataset_str): names = ['x',...

@zwhe99 Hi I am reaching out regarding if you see any sub-optimal behavior of DeepSpeed-fine-tuned model in comparison to non-DeepSpeed fine-tuned model. Especially the behavior that it stops generation after...

Hi Sean, thanks for your quick response! Are you using the deepspeed command line in the `train_dolly.py` script? My instance should be sufficiently big, both are p4.24dxlarge, so i assume...

Same issue here.