Skip to content

Update MoE functional test 2.0 configs #2731

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

Draft
Victarry wants to merge 14 commits into
base: dev
Choose a base branch
from
Draft

Update MoE functional test 2.0 configs #2731

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
14 commits
Select commit Hold shift + click to select a range
7c6908e
Fix model configs and add aux loss to logging.
Victarry Dec 19, 2025
9d84abc
Update golden values
Victarry Jan 6, 2026
4eec5e9
Formatting
Victarry Jan 6, 2026
0a97409
Add copyright.
Victarry Jan 6, 2026
395568c
Update golden values path.
Victarry Jan 6, 2026
6f0f30a
Add new test cases.
Victarry Jan 7, 2026
19ff70a
Update test cases and golden values.
Victarry Jan 7, 2026
26fe32a
Change NCCL Algo to Ring.
Victarry Jan 8, 2026
bef2a0e
Update golden values.
Victarry Jan 23, 2026
7021490
Merge branch 'dev' into denliu/moe2.0_functional_tests
Victarry Mar 1, 2026
b1016b3
Update golden values.
Victarry Mar 2, 2026
d19b63e
Fix moe2.0 recipe: wrap scalars in lists and move to h100/
Victarry Mar 2, 2026
2ac2741
Revert changes.
Victarry Mar 2, 2026
a57f8ff
Remove num-zeros from MoE 2.0 functional test METRICS
Victarry Mar 2, 2026
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
4 changes: 4 additions & 0 deletions tests/functional_tests/python_test_utils/get_test_results_from_tensorboard_logs.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,3 +1,4 @@
# Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
import os

os.environ["OPENBLAS_NUM_THREADS"] = "1"
Expand Down Expand Up @@ -63,6 +64,9 @@ def collect_train_test_metrics(
"lm loss",
"num-zeros",
"mtp_1 loss",
"load_balancing_loss",
"seq_load_balancing_loss",
"global_load_balancing_loss",
]
}

Expand Down
11 changes: 10 additions & 1 deletion tests/functional_tests/python_test_utils/test_pretraining_regular_pipeline.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
# Copyright (c) 2025 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
# Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.

import logging
from typing import Dict, List, Optional
Expand All @@ -19,6 +19,15 @@
"num-zeros": [common.DeterministicTest(), common.ApproximateTest(atol=0, rtol=0.05)],
"generated_tokens": [common.DeterministicTest(), common.ApproximateTest(atol=0, rtol=0.05)],
"logprobs": [common.DeterministicTest(), common.ApproximateTest(atol=0, rtol=0.05)],
"load_balancing_loss": [common.DeterministicTest(), common.ApproximateTest(atol=0, rtol=0.05)],
"seq_load_balancing_loss": [
common.DeterministicTest(),
common.ApproximateTest(atol=0, rtol=0.05),
],
"global_load_balancing_loss": [
common.DeterministicTest(),
common.ApproximateTest(atol=0, rtol=0.05),
],
}


Expand Down
Loading
Loading