Unverified Commit 46075b98 authored by Kirthi Shankar Sivamani's avatar Kirthi Shankar Sivamani Committed by GitHub
Browse files

[PyTorch] Fix distributed testing (#1219)



Fix
Signed-off-by: default avatarKirthi Shankar Sivamani <ksivamani@nvidia.com>
parent 728c558b
...@@ -4,16 +4,12 @@ ...@@ -4,16 +4,12 @@
set -e set -e
# pkg_resources is deprecated in setuptools 70+ and the packaging submodule
# has been removed from it. This is a temporary fix until upstream MLM fix.
pip install setuptools==69.5.1
: ${TE_PATH:=/opt/transformerengine} : ${TE_PATH:=/opt/transformerengine}
pytest -v -s $TE_PATH/tests/pytorch/distributed/test_comm_gemm_overlap.py pytest -v -s $TE_PATH/tests/pytorch/distributed/test_comm_gemm_overlap.py
pip install prettytable
git clone https://github.com/NVIDIA/Megatron-LM.git git clone https://github.com/NVIDIA/Megatron-LM.git
cd Megatron-LM cd Megatron-LM
git checkout bcce6f54e075e3c3374ea67adefe54f3f2da2b07 git checkout b3375a0e38c10e2300ef4be031f7dcabab52b448
sed -i -e '1504,1505d' megatron/model/transformer.py
pytest -v -s $TE_PATH/tests/pytorch/distributed/test_convergence.py pytest -v -s $TE_PATH/tests/pytorch/distributed/test_convergence.py
python $TE_PATH/tests/pytorch/distributed/print_logs.py python $TE_PATH/tests/pytorch/distributed/print_logs.py
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment