Tao Ma

**Describe the issue** I was trying to train dolly on my box. It runs fine when doing the following in Jupyter Notebook, but failed w/o detailed errors when run the...

[BUG] DeepSpeed tries to allocate memory from GPU 0 even though include was set to go with localhost:3,5

1

DeepSpeed (0.9.0) was started with `--include localhost:3,5` so to force the using of GPU 3 and GPU 5 for doing the training. But still the training failed because trying to...

**Describe the issue** I was trying to train dolly on my box. It runs fine when doing the following in Jupyter Notebook, but failed w/o detailed errors when run the...