Add PDL support for quant kernel and rope kernel (#9106)

2025-08-20 16:56:29 +08:00
parent c9bf3877a0
commit 42c8704560
7 changed files with 80 additions and 33 deletions
--- a/python/sglang/srt/server_args.py
+++ b/python/sglang/srt/server_args.py
@@ -550,7 +550,6 @@ class ServerArgs:
            assert (
                self.quantization == "modelopt_fp4"
            ), "modelopt_fp4 quantization is required for Flashinfer MOE"
-            os.environ["TRTLLM_ENABLE_PDL"] = "1"
            assert self.ep_size in [
                1,
                self.tp_size,