Skip to content

Bump to pytorch 25.06 and newer TE commit #14423

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Aug 11, 2025
Merged

Bump to pytorch 25.06 and newer TE commit #14423

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
2ea6816
Bump to pytorch 25.06 and newer TE commit
chtruong814 Aug 6, 2025
b2445ae
Fix mamba install
chtruong814 Aug 9, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion docker/common/install_dep.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -244,7 +244,7 @@ mcore() {

build() {
if [[ "${NVIDIA_PYTORCH_VERSION}" != "" ]]; then
pip wheel --no-deps --no-cache-dir --wheel-dir $WHEELS_DIR $MAMBA_DIR
pip wheel --no-deps --no-cache-dir --no-build-isolation --wheel-dir $WHEELS_DIR $MAMBA_DIR
Copy link
Collaborator Author

@chtruong814 chtruong814 Aug 10, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

A conversion test was failing due to mamba and this was how I had to fix it

  File "/home/TestData/HF_HOME/modules/transformers_modules/nemotronh-4B/modeling_nemotron_h.py", line 55, in <module>
    if is_mamba_2_ssm_available():
       ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.12/dist-packages/transformers/utils/import_utils.py", line 526, in is_mamba_2_ssm_available
    import mamba_ssm
  File "/usr/local/lib/python3.12/dist-packages/mamba_ssm/__init__.py", line 3, in <module>
    from mamba_ssm.ops.selective_scan_interface import selective_scan_fn, mamba_inner_fn
  File "/usr/local/lib/python3.12/dist-packages/mamba_ssm/ops/selective_scan_interface.py", line 18, in <module>
    import selective_scan_cuda
ImportError: /usr/local/lib/python3.12/dist-packages/selective_scan_cuda.cpython-312-x86_64-linux-gnu.so: undefined symbol: _ZN3c104cuda9SetDeviceEab

pip wheel --no-deps --no-cache-dir --wheel-dir $WHEELS_DIR $CAUSAL_CONV1D_DIR
fi

Expand Down
4 changes: 2 additions & 2 deletions requirements/manifest.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,13 +1,13 @@
{
"ngc-pytorch": "nvcr.io/nvidia/pytorch:25.05-py3",
"ngc-pytorch": "nvcr.io/nvidia/pytorch:25.06-py3",
"vcs-dependencies": {
"apex": {
"repo": "https://github.com/NVIDIA/Apex",
"ref": "810ffae374a2b9cb4b5c5e28eaeca7d7998fca0c"
},
"transformer_engine": {
"repo": "https://github.com/NVIDIA/TransformerEngine",
"ref": "f05f12c974b37c5bd8dfca3d2d294be53b66abfa"
"ref": "b97c2bf7b93be265fd4d60f600718920297049ef"
},
"megatron-lm": {
"repo": "https://github.com/NVIDIA/Megatron-LM",
Expand Down
Loading