search

ROCm / DeepSpeed

DeepSpeed is a deep learning optimization library that makes distributed training easy, efficient, and effective.
https://www.deepspeed.ai/
Apache License 2.0
4 stars 3 forks source link

IFU-master-2021-11-23 #48

Closed rraminen closed 2 years ago

rraminen commented 2 years ago

Resolved conflicts in setup.py, op_builder/init.py and pointed DeepSpeedExamples to the latest commit, 206e48b7638d1a36e90466071bd0f50844a3002b, after https://github.com/ROCmSoftwarePlatform/DeepSpeedExamples/pull/16 IFU

init.txt

setup.txt

I have verified Megatron-LM-v1.1.5 gpt2 and Bing BERT. Both work fine.

cc: @jithunnair-amd

jithunnair-amd commented 2 years ago

CI runs are reportedly still broken. And the GPT2 CI script needs to be updated to use the Megatron v1.1.5 script.