Skip to content

[https://nvbugs/7024242][fix] Scope the GLM-5.2 sm_103 tp8 accuracy engine's environment overrides #22032

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
trtllm-agent merged 1 commit into from
Nov 11, 2026
+5 −3
Merged

[https://nvbugs/7024242][fix] Scope the GLM-5.2 sm_103 tp8 accuracy engine's environment overrides #22032

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
7 changes: 5 additions & 2 deletions ...gration/defs/accuracy/modeling_v2/glm_moe_dsa/glm_5_2_nvfp4__sm_103__tp8/test_accuracy.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -35,6 +35,7 @@

from .....conftest import llm_models_root
from ....accuracy_core import GSM8K, LlmapiAccuracyTestHarness
from ..._env import scoped_env_overrides
from .._common import skip_not_sm103

# On every rank: check the target's step contract on the first forward, and run the
Expand All @@ -52,7 +53,6 @@ class TestModelingV2Glm52Nvfp4Sm103Tp8(LlmapiAccuracyTestHarness):
def _llm_kwargs(*, draft_len: int, kv_dtype: str, moe_backend: str, batch: int) -> dict:
return dict(
modeling_v2="require",
env_overrides=dict(_SM103_RANK_ENV),
speculative_config=MTPDecodingConfig(max_draft_len=draft_len),
kv_cache_config=KvCacheConfig(
dtype=kv_dtype, enable_block_reuse=False, free_gpu_memory_fraction=0.75
Expand All @@ -70,7 +70,10 @@ def _llm_kwargs(*, draft_len: int, kv_dtype: str, moe_backend: str, batch: int)
)

def _evaluate(self, **llm_kwargs) -> None:
with LLM(self.MODEL_PATH, **llm_kwargs) as llm:
with (
scoped_env_overrides(_SM103_RANK_ENV) as rank_env,
LLM(self.MODEL_PATH, env_overrides=rank_env, **llm_kwargs) as llm,
):
GSM8K(self.MODEL_NAME).evaluate(llm)

@skip_not_sm103
Expand Down
1 change: 0 additions & 1 deletion tests/integration/test_lists/waives.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -244,7 +244,6 @@ unittest/_torch/modeling_v2/models/minimax_m3/minimax_m3_nvfp4__sm_107__tp8/gemm
unittest/_torch/modeling_v2/models/minimax_m3/minimax_m3_nvfp4__sm_107__tp8/gemm/test_modeling_v2_m3_pinned_gemm.py::test_pinned_decision_matches_every_stock_candidate[1000.0-1536x6144-m16] SKIP (https://nvbugs/7054054)
unittest/_torch/modeling_v2/models/minimax_m3/minimax_m3_nvfp4__sm_107__tp8/gemm/test_modeling_v2_m3_pinned_gemm.py::test_pinned_decision_matches_every_stock_candidate[1000.0-1536x6144-m48] SKIP (https://nvbugs/7054054)
unittest/_torch/modeling_v2/models/minimax_m3/minimax_m3_nvfp4__sm_107__tp8/gemm/test_modeling_v2_m3_pinned_gemm.py::test_pinned_decision_matches_every_stock_candidate[1000.0-3072x6144-m24] SKIP (https://nvbugs/7054054)
unittest/_torch/modeling_v2/test_modeling_v2_env_scope.py SKIP (https://nvbugs/7024242)
unittest/_torch/modules/mamba/test_flashinfer_gdn_decode.py::test_fi_decode_misaligned_index_slice SKIP (https://nvbugs/6842831)
unittest/_torch/modules/tests_lora_modules/test_nemotron_h_lora_sanity.py::TestNemotronHLoRA::test_lora_pp2_sanity SKIP (https://nvbugs/6428124)
unittest/_torch/moe/test_dynamic_eplb_contracts.py::test_on_autotune_builds_balanced_powerlaw_input SKIP (https://nvbugs/7034332)
Expand Down
Loading