[Fix] Fix gpt oss triton kernels and upgrade flashinfer back to 0.6.11.post1 (#25335)

Co-authored-by: sglang-bot <sglang-bot@users.noreply.github.com>
Co-authored-by: b8zhong <b8zhong@users.noreply.github.com>
Co-authored-by: mmangkad <mmangkad@users.noreply.github.com>
This commit is contained in:
Baizhou Zhang
2026-05-15 01:04:56 -07:00
committed by GitHub
co-authored by sglang-bot b8zhong mmangkad
parent 4adfc6cf7e
commit 0c19540550
13 changed files with 147 additions and 53 deletions
+5 -5
View File
@@ -27,8 +27,8 @@ dependencies = [
"datasets",
"einops",
"fastapi",
"flashinfer_python==0.6.8.post1", # keep it aligned with jit-cache version in Dockerfile
"flashinfer_cubin==0.6.8.post1",
"flashinfer_python==0.6.11.post1", # keep it aligned with jit-cache version in Dockerfile
"flashinfer_cubin==0.6.11.post1",
"gguf",
"interegular",
"llguidance>=0.7.11,<0.8.0",
@@ -37,7 +37,7 @@ dependencies = [
"ninja",
"easydict", # Required by remote model code (e.g. DeepSeek-OCR) loaded via trust_remote_code; validated by transformers 5.4+ check_imports
"numpy",
"nvidia-cutlass-dsl==4.4.2",
"nvidia-cutlass-dsl==4.5.0",
"nvidia-ml-py",
"openai-harmony==0.0.4",
"openai==2.6.1",
@@ -53,14 +53,14 @@ dependencies = [
"pydantic",
"python-multipart",
"pyzmq>=25.1.2",
"quack-kernels>=0.3.0",
"quack-kernels>=0.4.1",
"requests",
"scipy",
"sentencepiece",
"setproctitle",
"flash-attn-4>=4.0.0b9",
"sgl-deep-gemm==0.1.0",
"sglang-kernel==0.4.2.post1",
"sglang-kernel==0.4.2.post2",
"soundfile==0.13.1",
"tiktoken",
"tilelang==0.1.8",