[watermarking] Dual-key gumbel-max watermarking for speculative decoding support (#56122)

Signed-off-by: Raphael Rialland <raphael.rialland@mistral.ai>
Signed-off-by: Simon Veitner <sveitner@redhat.com>
Signed-off-by: Tomas Ruiz <tomas.ruiz.te@gmail.com>
Co-authored-by: OpenAI Codex <codex@openai.com>
Co-authored-by: Simon Veitner <sveitner@redhat.com>
Co-authored-by: Tomas Ruiz <tomas.ruiz.te@gmail.com>
This commit is contained in:
Raphaël Rialland
2026-09-12 18:42:01 +00:00
committed by GitHub
co-authored by OpenAI Codex Simon Veitner Tomas Ruiz
parent 1ee4be4dbc
commit 7ee8a6dd01
28 changed files with 3460 additions and 245 deletions
@@ -44,6 +44,9 @@ from vllm.model_executor.warmup.replayssm_warmup import (
from vllm.model_executor.warmup.spec_decode_rejection_warmup import (
spec_decode_rejection_warmup,
)
from vllm.model_executor.warmup.watermark_sample_warmup import (
watermark_sample_warmup,
)
from vllm.platforms import current_platform
from vllm.utils.deep_gemm import is_deep_gemm_supported
from vllm.utils.flashinfer import has_flashinfer
@@ -207,6 +210,7 @@ def kernel_warmup(worker: "Worker", *, process_local_only: bool = False):
if enable_jit_warmup:
kimi_k3_triton_warmup(worker)
spec_decode_rejection_warmup(worker)
watermark_sample_warmup(worker)
qwen4_exp_qsa_triton_warmup(worker)
if enable_jit_warmup and current_platform.is_device_capability_family(100):