minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

[SPEC V2][2/N] feat: adaptive spec support spec v2 (#23336)

Browse Source 
Co-authored-by: Qiaolin-Yu <liin1211@outlook.com>
This commit is contained in:
shuwenn
2026-05-07 18:33:47 -07:00
committed by GitHub
co-authored by Qiaolin-Yu
parent 35870d55ac
commit d9dddd4d7d
12 changed files with 303 additions and 73 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test/registered/spec/eagle/test_adaptive_speculative.py
+26 -28
View File
@@ -6,7 +6,6 @@ from types import SimpleNamespace
import requests
from sglang.srt.environ import envs
from sglang.srt.utils import kill_process_tree
from sglang.test.ci.ci_register import register_cuda_ci
from sglang.test.run_eval import run_eval
@@ -59,33 +58,32 @@ class TestAdaptiveSpeculativeServer(CustomTestCase):
cls.adaptive_config_path = f.name
try:
with envs.SGLANG_ENABLE_SPEC_V2.override(False):
cls.process = popen_launch_server(
cls.model,
cls.base_url,
timeout=DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
other_args=[
"--trust-remote-code",
"--attention-backend",
"triton",
"--speculative-algorithm",
"EAGLE",
"--speculative-draft-model-path",
cls.draft_model,
"--speculative-num-steps",
"1",
"--speculative-eagle-topk",
"1",
"--speculative-num-draft-tokens",
"2",
"--speculative-adaptive",
"--speculative-adaptive-config",
cls.adaptive_config_path,
"--skip-server-warmup",
"--mem-fraction-static",
"0.7",
],
)
cls.process = popen_launch_server(
cls.model,
cls.base_url,
timeout=DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
other_args=[
"--trust-remote-code",
"--attention-backend",
"triton",
"--speculative-algorithm",
"EAGLE",
"--speculative-draft-model-path",
cls.draft_model,
"--speculative-num-steps",
"1",
"--speculative-eagle-topk",
"1",
"--speculative-num-draft-tokens",
"2",
"--speculative-adaptive",
"--speculative-adaptive-config",
cls.adaptive_config_path,
"--skip-server-warmup",
"--mem-fraction-static",
"0.7",
],
)
except Exception:
os.unlink(cls.adaptive_config_path)
raise
test/registered/unit/spec/test_adaptive_spec_params.py
+4 -3
View File
@@ -7,14 +7,15 @@ register_cpu_ci(est_time=6, suite="stage-a-test-cpu")
class TestAdaptiveSpeculativeParams(unittest.TestCase):
def test_initial_steps_snap_to_nearest_candidate_preferring_larger_step(self):
def test_initial_steps_added_to_candidates_when_missing(self):
params = AdaptiveSpeculativeParams(
initial_steps=2,
config={"candidate_steps": [1, 3, 7]},
)
self.assertEqual(params.current_steps, 3)
self.assertEqual(params.ema_accept_len, 2.0)
self.assertEqual(params.candidate_steps, [1, 2, 3, 7])
self.assertEqual(params.current_steps, 2)
self.assertEqual(params.ema_accept_len, 1.0)
def test_update_respects_warmup_and_interval(self):
params = AdaptiveSpeculativeParams(