Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
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: 2 additions & 0 deletions packages/sie_server/bundles/tensorrt-llm.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -17,6 +17,8 @@ deps:
torch: '==2.11.0'
transformers: '==5.5.4'
flashinfer-python: '==0.6.16'
# Keep the qualified build stable when the image resolver permits prereleases.
cuda-tile: '==1.6.0rc8'
triton: '==3.6.0'
nvidia-cutlass-dsl: '==4.5.0'
nvidia-cuda-tileiras: '==13.1.80'
Expand Down
6 changes: 6 additions & 0 deletions packages/sie_server/tests/config/test_bundle_coverage.py
Original file line number Diff line number Diff line change
Expand Up @@ -123,6 +123,12 @@ def test_bundle_and_model_dirs_are_non_empty() -> None:
assert sorted(MODELS_DIR.glob("*.yaml")), f"No model YAML files found in {MODELS_DIR}"


def test_tensorrt_llm_bundle_pins_qualified_cuda_tile() -> None:
bundle = yaml.safe_load((BUNDLES_DIR / "tensorrt-llm.yaml").read_text()) or {}

assert bundle["deps"]["cuda-tile"] == "==1.6.0rc8"


def test_candle_bundle_only_exposes_profile_variants() -> None:
matches = match_bundle_models(BUNDLES_DIR / "candle.yaml", MODELS_DIR)
assert matches, "candle bundle should expose at least one model profile variant"
Expand Down
Loading