Skip to content

add Flashinfer sampling backend #1328

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
blueswhen wants to merge 2 commits into
base: main
Choose a base branch
from
Open

add Flashinfer sampling backend #1328

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
a687cbe
feat: add flashinfer sampling_backend
Jun 2, 2026
97a42e9
feat: remove sglang sampling_backend
Jun 2, 2026
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
4 changes: 2 additions & 2 deletions lightllm/server/api_cli.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -644,10 +644,10 @@ def make_argument_parser() -> argparse.ArgumentParser:
parser.add_argument(
"--sampling_backend",
type=str,
choices=["triton", "sglang_kernel"],
choices=["triton", "flashinfer"],
default="triton",
help="""sampling used impl. 'triton' is use torch and triton kernel,
sglang_kernel use sglang_kernel impl""",
flashinfer use flashinfer sampling impl""",
)
parser.add_argument(
"--penalty_counter_mode",
Expand Down
2 changes: 1 addition & 1 deletion lightllm/server/core/objs/start_args_type.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -149,7 +149,7 @@ class StartArgs:
default="None", metadata={"choices": ["None", "int8kv", "int4kv", "fp8kv_sph", "fp8kv_spt", "fp8kv_dsa"]}
)
llm_kv_quant_group_size: int = field(default=8)
sampling_backend: str = field(default="triton", metadata={"choices": ["triton", "sglang_kernel"]})
sampling_backend: str = field(default="triton", metadata={"choices": ["triton", "flashinfer"]})
penalty_counter_mode: str = field(
default="gpu_counter", metadata={"choices": ["cpu_counter", "pin_mem_counter", "gpu_counter"]}
)
Expand Down
8 changes: 5 additions & 3 deletions lightllm/server/router/model_infer/mode_backend/generic_post_process.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -114,7 +114,9 @@ def _top_p_top_k_sample(
b_top_ks: torch.Tensor,
exist_req_use_random_seed: bool,
) -> Tuple[torch.Tensor, torch.Tensor]:
if get_env_start_args().sampling_backend == "triton":
sampling_backend = get_env_start_args().sampling_backend

if sampling_backend == "triton":
Copy link
Copy Markdown
Contributor

@gemini-code-assist gemini-code-assist Bot Jun 2, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

medium

The flashinfer sampling backend does not support per-request random seeds (exist_req_use_random_seed), which will break deterministic/seeded generation when a custom seed is provided. To ensure correctness, we should fall back to the triton sampling implementation when exist_req_use_random_seed is True.

Suggested change
if sampling_backend == "triton":
if sampling_backend == "triton" or exist_req_use_random_seed:

probs_sort, probs_idx = _top_p_top_k(probs, b_top_ps, b_top_ks)
if not exist_req_use_random_seed:
sampled_index = torch.multinomial(probs_sort, num_samples=1, replacement=True)
Expand All @@ -124,8 +126,8 @@ def _top_p_top_k_sample(
next_token_logprobs = torch.log(torch.gather(probs_sort, dim=1, index=sampled_index))
return next_token_ids.view(-1), next_token_logprobs.view(-1)

elif get_env_start_args().sampling_backend == "sglang_kernel":
from sgl_kernel import top_k_top_p_sampling_from_probs
elif sampling_backend == "flashinfer":
from flashinfer.sampling import top_k_top_p_sampling_from_probs

batch_next_token_ids = top_k_top_p_sampling_from_probs(
probs,
Expand Down
4 changes: 2 additions & 2 deletions requirements.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -80,8 +80,8 @@ frozendict==2.4.6
atomics==1.0.3
easydict==1.13
hypercorn==0.18.0
flashinfer-python==0.6.8.post1
flashinfer-cubin==0.6.8.post1
flashinfer-python==0.6.12
flashinfer-cubin==0.6.12
sglang-kernel==0.4.2.post1
httpx==0.28.1
librosa==0.11.0
Expand Down
14 changes: 12 additions & 2 deletions test/benchmark/service/benchmark_qps.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -108,6 +108,11 @@ def get_custom_input_data(data_path, output_len, tokenizer, range_ratio):


model_name = []
sampling_config = {
"temperature": 1.0,
"top_p": 0.9,
"top_k": -1,
}


# Minimal fix: one retry on transient network errors.
Expand All @@ -123,7 +128,9 @@ async def async_post_stream_openai(url, prompt, max_new_tokens, session):
"max_tokens": max_new_tokens,
"ignore_eos": True,
"stream": True,
"temperature": 0.0,
"temperature": sampling_config["temperature"],
"top_p": sampling_config["top_p"],
"top_k": sampling_config["top_k"],
"best_of": 1,
}
headers = {"Content-Type": "application/json"}
Expand Down Expand Up @@ -166,9 +173,12 @@ async def async_post_stream_lightllm(url, prompt, max_new_tokens, session):
data = {
"inputs": text_input,
"parameters": {
"do_sample": False,
"do_sample": True,
"ignore_eos": True,
"max_new_tokens": max_new_tokens,
"temperature": sampling_config["temperature"],
"top_p": sampling_config["top_p"],
"top_k": sampling_config["top_k"],
"add_special_tokens": False,
},
}
Expand Down
Loading