Files

T

d229952e25 [Fix] Preserve model runner contracts in prefill CUDA graphs (#35452 )

Co-authored-by: Oasis-Git <ayw.sirius19@gmail.com>
Co-authored-by: Ke Bao <ispobaoke@gmail.com>

2026-09-20 13:53:14 -07:00

benchmark

[Cleanup] Deduplicate kernel tests, diffusion fixtures and benchmark helpers (#40265 )

2026-09-19 19:45:38 +08:00

cli

[diffusion] feat: out of tree platform support (#37547 )

2026-09-19 09:35:58 +08:00

kernels

Add out-of-tree DFlash extension points (#38740 )

2026-09-20 14:54:46 +08:00

lang

[Config] msgspec.Struct for the config tier (#38753 )

2026-09-09 19:41:19 -07:00

multimodal_gen

[Diffusion] Enable shared RMSNorm dispatch for SenseNova-U1 (#39705 )

2026-09-20 17:08:02 +03:00

srt

[Fix] Preserve model runner contracts in prefill CUDA graphs (#35452 )

2026-09-20 13:53:14 -07:00

test

Update test cases and performance testing framework (#40392 )

2026-09-20 22:42:00 +08:00

__init__.py

[Misc] Clean up python/sglang package structure (#35062 )

2026-08-17 14:24:35 -07:00

_platform_stubs.py

[Fix] Register triton.runtime.cache.triton_key in the MPS stub so torch.compile keeps working (#37937 )

2026-09-03 22:56:14 -07:00

bench_offline_throughput.py

…

bench_one_batch_server.py

…

bench_one_batch.py

…

bench_serving.py

…

check_env.py

Remove the torchao integration (--torchao-config) (#34304 )

2026-08-14 21:49:11 +08:00

compile_deep_gemm.py

config: a parallel size has one spelling; a patched scope declares its own (#36621 )

2026-08-27 12:56:42 -07:00

launch_server.py

[Fix] Seed raw tokenizer_path for smg-grpc-servicer in gRPC mode (#39105 )

2026-09-12 14:41:15 +08:00

profiler.py

…

README.md

[Misc] Clean up python/sglang package structure (#35062 )

2026-08-17 14:24:35 -07:00

utils.py

[diffusion] fix: fix local-path detection for MiniMax-H3 and other non-diffusers models (#33365 )

2026-08-04 14:03:35 +08:00

version.py

…

README.md

Code Structure

Folders

benchmark/: Benchmark implementations and dataset utilities.
cli/: Command-line interface commands and entrypoints.
kernels/: Kernel interfaces, implementations, selection, and debugging utilities shared by the runtimes.
lang/: Deprecated language frontend that is no longer actively maintained.
multimodal_gen/: Core runtime for image, video, and audio generation models, most of which are diffusion models.
srt/: Core runtime for autoregressive language models. (SRT = SGLang Runtime.)
test/: Shared test and evaluation utilities.

Files

README.md: This package structure overview.
__init__.py: Package initialization and public Python APIs.
bench_offline_throughput.py: Deprecated wrapper for the offline throughput benchmark.
bench_one_batch.py: Deprecated wrapper for the one-batch benchmark.
bench_one_batch_server.py: Deprecated wrapper for the server-based one-batch benchmark.
bench_serving.py: Deprecated wrapper for the online serving benchmark.
check_env.py: Environment and dependency diagnostics.
compile_deep_gemm.py: DeepGEMM kernel precompilation entrypoint.
launch_server.py: Compatibility entrypoint for launching an inference server.
profiler.py: Client entrypoint for collecting server profiling traces.
utils.py: Common package utilities.
version.py: Public package version resolution.