diff --git a/studio/backend/core/inference/llama_cpp.py b/studio/backend/core/inference/llama_cpp.py
index 260e675a73..bd55346f4b 100644
--- a/studio/backend/core/inference/llama_cpp.py
+++ b/studio/backend/core/inference/llama_cpp.py
@@ -522,11 +522,50 @@ def _build_ngram_mod_flags(
return []
-# Canonical Speculative Decoding modes exposed by the Studio chat UI.
-# The dropdown renders five options (auto, mtp, ngram, mtp+ngram, off);
-# the load API also accepts legacy values that the original Switch and
-# external callers emit (default, draft-mtp, ngram-mod, ngram-simple).
-_CANONICAL_SPEC_MODES = {"auto", "mtp", "ngram", "mtp+ngram", "off", "ngram-simple"}
+def _build_ngram_map_k_flags(
+ caps: Optional[dict],
+ *,
+ variant: str,
+ size_n: int = 16,
+ size_m: int = 24,
+ min_hits: int = 1,
+) -> list[str]:
+ """Emit knob flags for ``ngram-map-k`` / ``ngram-map-k4v`` if the
+ bundled llama-server advertises them (added upstream in #23269 and
+ sibling commits). Returns ``[]`` when the binary doesn't.
+ """
+ cap_key = (
+ "supports_ngram_map_k4v"
+ if variant == "ngram-map-k4v"
+ else "supports_ngram_map_k"
+ )
+ if not (caps and caps.get(cap_key)):
+ return []
+ prefix = f"--spec-{variant}"
+ return [
+ f"{prefix}-size-n",
+ str(size_n),
+ f"{prefix}-size-m",
+ str(size_m),
+ f"{prefix}-min-hits",
+ str(min_hits),
+ ]
+
+
+# Canonical Speculative Decoding modes exposed by the Studio chat UI
+# (5-mode dropdown) plus power-user n-gram variants accepted via the
+# load API. ``ngram-map-k`` / ``ngram-map-k4v`` are not in the dropdown
+# but the resolver emits them correctly when sent via API.
+_CANONICAL_SPEC_MODES = {
+ "auto",
+ "mtp",
+ "ngram",
+ "mtp+ngram",
+ "off",
+ "ngram-simple",
+ "ngram-map-k",
+ "ngram-map-k4v",
+}
_LEGACY_SPEC_MODE_MAP = {
"default": "auto",
"draft-mtp": "mtp",
@@ -538,9 +577,10 @@ def _canonicalize_spec_mode(value):
"""Map any accepted ``speculative_type`` input onto a canonical mode.
Returns one of ``auto``, ``mtp``, ``ngram``, ``mtp+ngram``, ``off``,
- ``ngram-simple``, or ``None`` (callers treat ``None`` as ``auto``).
- Unknown strings collapse to ``auto`` so a stale UI value or typo
- falls back to the safe platform-aware path.
+ ``ngram-simple``, ``ngram-map-k``, ``ngram-map-k4v``, or ``None``
+ (callers treat ``None`` as ``auto``). Unknown strings collapse to
+ ``auto`` so a stale UI value or typo falls back to the safe
+ platform-aware path.
"""
if value is None:
return None
@@ -641,6 +681,10 @@ def __init__(self):
self._requested_spec_mode: Optional[str] = None
# User-supplied --spec-draft-n-max override (None = platform default).
self._spec_draft_n_max: Optional[int] = None
+ # User-supplied --spec-draft-p-min override (None = platform default).
+ # Only emitted for MTP / MTP+Ngram modes; the flag became functional
+ # in llama.cpp #23269 (default upstream dropped to 0.0).
+ self._spec_draft_p_min: Optional[float] = None
# KV-cache estimation fields (populated by _read_gguf_metadata)
self._n_layers: Optional[int] = None
self._n_kv_heads: Optional[int] = None
@@ -932,6 +976,12 @@ def spec_draft_n_max(self) -> Optional[int]:
when the platform default (6 GPU / 3 CPU) is in effect."""
return self._spec_draft_n_max
+ @property
+ def spec_draft_p_min(self) -> Optional[float]:
+ """User --spec-draft-p-min override on the load, or None when
+ the llama-server default (0.0 since #23269) is in effect."""
+ return self._spec_draft_p_min
+
# ── Binary discovery ──────────────────────────────────────────
@staticmethod
@@ -1059,7 +1109,8 @@ def probe_server_capabilities(
) -> dict[str, object]:
"""Parse `llama-server --help` for feature flags. Returns
{found, mtp_token, supports_mtp, ngram_mod_flavor,
- supports_ngram_mod, spec_draft_n_max_flag}.
+ supports_ngram_mod, spec_draft_n_max_flag, spec_draft_p_min_flag,
+ supports_ngram_map_k, supports_ngram_map_k4v}.
``ngram_mod_flavor`` is ``"new"`` when the binary exposes the
post-rename ``--spec-ngram-mod-n-match / -n-min / -n-max`` as
@@ -1072,6 +1123,13 @@ def probe_server_capabilities(
``spec_draft_n_max_flag`` is the actual flag name the binary
accepts: ``--spec-draft-n-max`` on post-rename builds, or
``--draft-max`` on legacy. ``None`` means n_max cannot be set.
+
+ ``spec_draft_p_min_flag`` is the flag name when the binary
+ advertises ``--spec-draft-p-min`` (functional from llama.cpp
+ #23269 onward; before that it was a knob with no effect on
+ MTP). ``supports_ngram_map_k`` / ``supports_ngram_map_k4v``
+ signal that the binary accepts these power-user spec types
+ with their own knob sets.
"""
bin_path = binary or cls._find_llama_server_binary()
if not bin_path or not Path(bin_path).is_file():
@@ -1082,6 +1140,9 @@ def probe_server_capabilities(
"ngram_mod_flavor": None,
"supports_ngram_mod": False,
"spec_draft_n_max_flag": None,
+ "spec_draft_p_min_flag": None,
+ "supports_ngram_map_k": False,
+ "supports_ngram_map_k4v": False,
}
try:
mtime = int(Path(bin_path).stat().st_mtime)
@@ -1095,6 +1156,9 @@ def probe_server_capabilities(
mtp_token: Optional[str] = None
ngram_mod_flavor: Optional[str] = None
spec_draft_n_max_flag: Optional[str] = None
+ spec_draft_p_min_flag: Optional[str] = None
+ supports_ngram_map_k = False
+ supports_ngram_map_k4v = False
try:
result = subprocess.run(
[bin_path, "--help"],
@@ -1184,6 +1248,24 @@ def _is_real(flag: str) -> bool:
spec_draft_n_max_flag = "--spec-draft-n-max"
elif _is_real("--draft-max"):
spec_draft_n_max_flag = "--draft-max"
+
+ # p_min flag: present on builds that include the MTP cleanup
+ # (llama.cpp #23269) or earlier draft-model spec impls.
+ if _is_real("--spec-draft-p-min"):
+ spec_draft_p_min_flag = "--spec-draft-p-min"
+
+ # ngram-map-k / ngram-map-k4v: power-user spec types added
+ # alongside ngram-mod. Each carries its own knob triplet.
+ supports_ngram_map_k = (
+ _is_real("--spec-ngram-map-k-size-n")
+ and _is_real("--spec-ngram-map-k-size-m")
+ and _is_real("--spec-ngram-map-k-min-hits")
+ )
+ supports_ngram_map_k4v = (
+ _is_real("--spec-ngram-map-k4v-size-n")
+ and _is_real("--spec-ngram-map-k4v-size-m")
+ and _is_real("--spec-ngram-map-k4v-min-hits")
+ )
except (OSError, subprocess.SubprocessError) as exc:
logger.debug(f"llama-server --help probe failed: {exc}")
@@ -1194,6 +1276,9 @@ def _is_real(flag: str) -> bool:
"ngram_mod_flavor": ngram_mod_flavor,
"supports_ngram_mod": ngram_mod_flavor is not None,
"spec_draft_n_max_flag": spec_draft_n_max_flag,
+ "spec_draft_p_min_flag": spec_draft_p_min_flag,
+ "supports_ngram_map_k": supports_ngram_map_k,
+ "supports_ngram_map_k4v": supports_ngram_map_k4v,
}
cls._capability_cache[cache_key] = info
return info
@@ -2502,6 +2587,7 @@ def load_model(
cache_type_kv: Optional[str] = None,
speculative_type: Optional[str] = None,
spec_draft_n_max: Optional[int] = None,
+ spec_draft_p_min: Optional[float] = None,
n_threads: Optional[int] = None,
n_gpu_layers: Optional[int] = None, # Accepted for caller compat, unused
n_parallel: int = 1,
@@ -2534,6 +2620,7 @@ def load_model(
cache_type_kv = cache_type_kv,
speculative_type = speculative_type,
spec_draft_n_max = spec_draft_n_max,
+ spec_draft_p_min = spec_draft_p_min,
chat_template_override = chat_template_override,
extra_args = extra_args,
is_vision = is_vision,
@@ -2915,6 +3002,7 @@ def load_model(
spec_flags = self._build_speculative_flags(
speculative_type = speculative_type,
spec_draft_n_max = spec_draft_n_max,
+ spec_draft_p_min = spec_draft_p_min,
extra_args = extra_args,
model_identifier = model_identifier,
model_path = model_path,
@@ -3258,6 +3346,7 @@ def _build_speculative_flags(
*,
speculative_type: Optional[str],
spec_draft_n_max: Optional[int],
+ spec_draft_p_min: Optional[float] = None,
extra_args: Optional[List[str]],
model_identifier: str,
model_path: Optional[str],
@@ -3299,6 +3388,7 @@ def _build_speculative_flags(
flags: List[str] = []
# Reset; emit branches re-set on the resolved emission.
self._spec_draft_n_max = None
+ self._spec_draft_p_min = None
self._speculative_type = None
# Canonical UI-facing requested mode: auto / mtp / ngram /
@@ -3334,6 +3424,22 @@ def _resolved_draft_n_max() -> int:
return n
return 2 if gpus else 3
+ def _maybe_emit_p_min(caps: Optional[dict]) -> None:
+ """Append --spec-draft-p-min if the user supplied a value AND
+ the binary advertises the flag (functional from #23269)."""
+ if spec_draft_p_min is None:
+ return
+ p_flag = (caps or {}).get("spec_draft_p_min_flag")
+ if not p_flag:
+ logger.warning(
+ "llama-server lacks --spec-draft-p-min; ignoring "
+ "spec_draft_p_min override. Run `unsloth studio update`."
+ )
+ return
+ p_min = float(spec_draft_p_min)
+ self._spec_draft_p_min = p_min
+ flags.extend([p_flag, f"{p_min:.4f}"])
+
def _emit_mtp(*, chain_ngram: bool) -> bool:
"""Append --spec-type mtp[/draft-mtp][,ngram-mod] + n-max."""
caps = self.probe_server_capabilities(binary)
@@ -3376,6 +3482,7 @@ def _emit_mtp(*, chain_ngram: bool) -> bool:
str(draft_n_max),
]
)
+ _maybe_emit_p_min(caps)
self._speculative_type = "draft-mtp"
chain_label = "chained ngram-mod" if chain_ngram else "MTP-only"
logger.info(f"Spec decoding: {mtp_token} ({chain_label})")
@@ -3405,6 +3512,25 @@ def _emit_ngram_mod() -> bool:
if effective_mode == "ngram":
_emit_ngram_mod()
return flags
+ if effective_mode in ("ngram-map-k", "ngram-map-k4v"):
+ map_caps = self.probe_server_capabilities(binary)
+ supported = map_caps.get(
+ "supports_ngram_map_k4v"
+ if effective_mode == "ngram-map-k4v"
+ else "supports_ngram_map_k"
+ )
+ if not supported:
+ logger.warning(
+ f"llama-server lacks --spec-type {effective_mode}; "
+ "run `unsloth studio update`. Loading without "
+ "speculative decoding."
+ )
+ return flags
+ flags.extend(["--spec-type", effective_mode])
+ flags.extend(_build_ngram_map_k_flags(map_caps, variant = effective_mode))
+ self._speculative_type = effective_mode
+ logger.info(f"Spec decoding: {effective_mode}")
+ return flags
if effective_mode == "mtp":
if _mtp_too_small:
logger.warning(
@@ -3480,6 +3606,7 @@ def _already_in_target_state(
is_vision: bool,
gguf_path: Optional[str] = None,
spec_draft_n_max: Optional[int] = None,
+ spec_draft_p_min: Optional[float] = None,
) -> bool:
"""True iff the live server already satisfies these load kwargs.
@@ -3540,6 +3667,16 @@ def _norm(value):
):
return False
+ # Same treatment for spec_draft_p_min: only relevant when MTP is
+ # the resolved emit. None on either side means "default" (0.0
+ # since llama.cpp #23269) and matches anything.
+ if (
+ self._speculative_type == "draft-mtp"
+ and spec_draft_p_min is not None
+ and abs(float(spec_draft_p_min) - (self._spec_draft_p_min or 0.0)) > 1e-6
+ ):
+ return False
+
if (self._chat_template_override or None) != (chat_template_override or None):
return False
@@ -3608,6 +3745,7 @@ def unload_model(self) -> bool:
self._speculative_type = None
self._requested_spec_mode = None
self._spec_draft_n_max = None
+ self._spec_draft_p_min = None
self._n_layers = None
self._n_kv_heads = None
self._n_kv_heads_by_layer = None
diff --git a/studio/backend/models/inference.py b/studio/backend/models/inference.py
index f8bf219846..5ccb2ab447 100644
--- a/studio/backend/models/inference.py
+++ b/studio/backend/models/inference.py
@@ -71,14 +71,16 @@ def normalize_blank_chat_template_override(
speculative_type: Optional[str] = Field(
None,
description = (
- "Speculative decoding mode for GGUF models. Canonical values: "
- "'auto' (platform-aware: MTP on MTP GGUFs, ngram-mod fallback "
- "for sub-3B), 'mtp' (force draft-mtp only on both GPU and CPU), "
- "'ngram' (force ngram-mod only), 'mtp+ngram' (force "
- "ngram-mod+draft-mtp chain on both platforms), 'off' (disabled). "
- "Legacy values 'default' (-> auto), 'draft-mtp' (-> mtp), "
- "'ngram-mod' (-> ngram), and 'ngram-simple' (kept as-is) are "
- "still accepted. Ignored for non-GGUF and vision models."
+ "Speculative decoding mode for GGUF models. Canonical values "
+ "exposed by the Studio dropdown: 'auto' (platform-aware: MTP "
+ "on MTP GGUFs, ngram-mod fallback for sub-3B), 'mtp' (force "
+ "draft-mtp only on both GPU and CPU), 'ngram' (force "
+ "ngram-mod only), 'mtp+ngram' (force ngram-mod+draft-mtp "
+ "chain on both platforms), 'off' (disabled). Power-user "
+ "spec types accepted via API: 'ngram-simple', 'ngram-map-k', "
+ "'ngram-map-k4v'. Legacy values 'default' (-> auto), "
+ "'draft-mtp' (-> mtp), 'ngram-mod' (-> ngram) are still "
+ "accepted. Ignored for non-GGUF and vision models."
),
)
spec_draft_n_max: Optional[int] = Field(
@@ -93,6 +95,19 @@ def normalize_blank_chat_template_override(
"'mtp' or 'mtp+ngram'."
),
)
+ spec_draft_p_min: Optional[float] = Field(
+ None,
+ ge = 0.0,
+ le = 1.0,
+ description = (
+ "Min draft probability for MTP speculative decoding "
+ "(--spec-draft-p-min). Drafts with predicted probability "
+ "below this threshold are rejected. Defaults to 0.0 (no "
+ "filtering) since llama.cpp #23269; before that the flag "
+ "existed but was non-functional. Only applied when "
+ "speculative_type resolves to 'mtp' or 'mtp+ngram'."
+ ),
+ )
llama_extra_args: Optional[List[str]] = Field(
None,
description = (
@@ -242,8 +257,9 @@ class LoadResponse(BaseModel):
description = (
"Canonical UI-facing requested speculative decoding mode "
"('auto' / 'mtp' / 'ngram' / 'mtp+ngram' / 'off' / "
- "'ngram-simple'), round-tripped from the original LoadRequest "
- "via _canonicalize_spec_mode. None when no model is loaded."
+ "'ngram-simple' / 'ngram-map-k' / 'ngram-map-k4v'), "
+ "round-tripped from the original LoadRequest via "
+ "_canonicalize_spec_mode. None when no model is loaded."
),
)
spec_draft_n_max: Optional[int] = Field(
@@ -253,6 +269,13 @@ class LoadResponse(BaseModel):
"None when the platform default is in effect."
),
)
+ spec_draft_p_min: Optional[float] = Field(
+ None,
+ description = (
+ "Active --spec-draft-p-min for MTP speculative decoding, or "
+ "None when the llama-server default (0.0) is in effect."
+ ),
+ )
class UnloadResponse(BaseModel):
@@ -376,8 +399,9 @@ class InferenceStatusResponse(BaseModel):
description = (
"Canonical UI-facing requested speculative decoding mode "
"('auto' / 'mtp' / 'ngram' / 'mtp+ngram' / 'off' / "
- "'ngram-simple'), round-tripped from the original LoadRequest. "
- "None when no model is loaded."
+ "'ngram-simple' / 'ngram-map-k' / 'ngram-map-k4v'), "
+ "round-tripped from the original LoadRequest. None when no "
+ "model is loaded."
),
)
spec_draft_n_max: Optional[int] = Field(
@@ -387,6 +411,13 @@ class InferenceStatusResponse(BaseModel):
"None when the platform default is in effect."
),
)
+ spec_draft_p_min: Optional[float] = Field(
+ None,
+ description = (
+ "Active --spec-draft-p-min for MTP speculative decoding, or "
+ "None when the llama-server default (0.0) is in effect."
+ ),
+ )
llama_cpp_supports_mtp: bool = Field(
True,
description = (
diff --git a/studio/backend/routes/inference.py b/studio/backend/routes/inference.py
index cde86998d2..aec753f0e2 100644
--- a/studio/backend/routes/inference.py
+++ b/studio/backend/routes/inference.py
@@ -504,11 +504,21 @@ def _request_matches_loaded_settings(
backend_mode = llama_backend.requested_spec_mode or "auto"
if req_mode != backend_mode:
return False
- # spec_draft_n_max only matters when an MTP variant is engaged; None
- # means "platform default" and matches whatever the backend chose.
+ # spec_draft_n_max / spec_draft_p_min only matter when an MTP variant
+ # is engaged; None on the request means "platform default" and
+ # matches whatever the backend chose.
if backend_mode in ("mtp", "mtp+ngram") and request.spec_draft_n_max is not None:
if int(request.spec_draft_n_max) != (llama_backend.spec_draft_n_max or 0):
return False
+ if (
+ backend_mode in ("mtp", "mtp+ngram")
+ and request.spec_draft_p_min is not None
+ and abs(
+ float(request.spec_draft_p_min) - (llama_backend.spec_draft_p_min or 0.0)
+ )
+ > 1e-6
+ ):
+ return False
if (request.chat_template_override or None) != (
llama_backend.chat_template_override or None
):
@@ -649,6 +659,7 @@ async def load_model(
chat_template = llama_backend.chat_template,
speculative_type = llama_backend.requested_spec_mode,
spec_draft_n_max = llama_backend.spec_draft_n_max,
+ spec_draft_p_min = llama_backend.spec_draft_p_min,
)
else:
if (
@@ -780,6 +791,7 @@ async def load_model(
strip_spec = (
"speculative_type" in fields_set
or "spec_draft_n_max" in fields_set
+ or "spec_draft_p_min" in fields_set
),
strip_template = "chat_template_override" in fields_set,
)
@@ -822,6 +834,7 @@ async def load_model(
cache_type_kv = request.cache_type_kv,
speculative_type = request.speculative_type,
spec_draft_n_max = request.spec_draft_n_max,
+ spec_draft_p_min = request.spec_draft_p_min,
n_parallel = _n_parallel,
extra_args = extra_llama_args,
)
@@ -846,6 +859,7 @@ async def load_model(
cache_type_kv = request.cache_type_kv,
speculative_type = request.speculative_type,
spec_draft_n_max = request.spec_draft_n_max,
+ spec_draft_p_min = request.spec_draft_p_min,
n_parallel = _n_parallel,
extra_args = extra_llama_args,
)
@@ -906,6 +920,7 @@ async def load_model(
chat_template = llama_backend.chat_template,
speculative_type = llama_backend.requested_spec_mode,
spec_draft_n_max = llama_backend.spec_draft_n_max,
+ spec_draft_p_min = llama_backend.spec_draft_p_min,
)
# ── Standard path: load via Unsloth/transformers ──────────
@@ -1395,6 +1410,7 @@ async def get_status(
chat_template_override = llama_backend.chat_template_override,
speculative_type = llama_backend.requested_spec_mode,
spec_draft_n_max = llama_backend.spec_draft_n_max,
+ spec_draft_p_min = llama_backend.spec_draft_p_min,
llama_cpp_supports_mtp = _supports_mtp,
llama_cpp_prebuilt_stale = _stale,
llama_cpp_installed_tag = _installed_tag,
diff --git a/studio/backend/tests/test_llama_cpp_mtp_detection.py b/studio/backend/tests/test_llama_cpp_mtp_detection.py
index 4a8276adc0..a6b0f9a9f9 100644
--- a/studio/backend/tests/test_llama_cpp_mtp_detection.py
+++ b/studio/backend/tests/test_llama_cpp_mtp_detection.py
@@ -1011,7 +1011,15 @@ def test_canonicalize_spec_mode(value, expected):
# ── _build_speculative_flags resolver matrix ──────────────────────
-def _resolver_backend(monkeypatch, *, ngram_supported = True, mtp_token = "draft-mtp"):
+def _resolver_backend(
+ monkeypatch,
+ *,
+ ngram_supported = True,
+ mtp_token = "draft-mtp",
+ p_min_flag = "--spec-draft-p-min",
+ supports_ngram_map_k = True,
+ supports_ngram_map_k4v = True,
+):
"""Backend with a deterministic probe so the resolver is hermetic."""
fake = {
"found": True,
@@ -1020,6 +1028,9 @@ def _resolver_backend(monkeypatch, *, ngram_supported = True, mtp_token = "draft
"ngram_mod_flavor": "new" if ngram_supported else None,
"supports_ngram_mod": bool(ngram_supported),
"spec_draft_n_max_flag": "--spec-draft-n-max",
+ "spec_draft_p_min_flag": p_min_flag,
+ "supports_ngram_map_k": supports_ngram_map_k,
+ "supports_ngram_map_k4v": supports_ngram_map_k4v,
}
monkeypatch.setattr(
LlamaCppBackend,
@@ -1198,3 +1209,132 @@ def test_build_speculative_flags_mtp_token_missing_logs_and_skips(monkeypatch):
# _requested_spec_mode still reflects the user's choice.
assert backend.requested_spec_mode == "mtp"
assert backend.speculative_type is None
+
+
+# ── spec_draft_p_min ────────────────────────────────────────────
+
+
+@pytest.mark.parametrize("mode", ["mtp", "mtp+ngram"])
+def test_build_speculative_flags_emits_p_min_for_mtp_modes(monkeypatch, mode):
+ backend = _resolver_backend(monkeypatch)
+ flags = backend._build_speculative_flags(
+ speculative_type = mode,
+ spec_draft_n_max = None,
+ spec_draft_p_min = 0.25,
+ extra_args = None,
+ model_identifier = _MTP_MODEL,
+ model_path = None,
+ gpus = True,
+ binary = "/fake/llama-server",
+ )
+ parsed = _flags_dict(flags)
+ assert parsed.get("--spec-draft-p-min") == "0.2500"
+ assert backend.spec_draft_p_min == 0.25
+
+
+@pytest.mark.parametrize("mode", ["auto", "ngram", "off", "ngram-simple"])
+def test_build_speculative_flags_does_not_emit_p_min_for_non_mtp_modes(
+ monkeypatch, mode
+):
+ backend = _resolver_backend(monkeypatch)
+ flags = backend._build_speculative_flags(
+ speculative_type = mode,
+ spec_draft_n_max = None,
+ spec_draft_p_min = 0.25,
+ extra_args = None,
+ # Use a non-MTP model so "auto" doesn't promote to draft-mtp
+ # and accidentally pick up the p-min emission.
+ model_identifier = _NON_MTP_MODEL,
+ model_path = None,
+ gpus = True,
+ binary = "/fake/llama-server",
+ )
+ assert "--spec-draft-p-min" not in flags
+ assert backend.spec_draft_p_min is None
+
+
+def test_build_speculative_flags_p_min_skipped_when_binary_lacks_flag(monkeypatch):
+ # Older llama-server (no #23269): p_min flag is None in the probe.
+ # Resolver should log a warning and skip emission rather than crash.
+ backend = _resolver_backend(monkeypatch, p_min_flag = None)
+ flags = backend._build_speculative_flags(
+ speculative_type = "mtp",
+ spec_draft_n_max = None,
+ spec_draft_p_min = 0.5,
+ extra_args = None,
+ model_identifier = _MTP_MODEL,
+ model_path = None,
+ gpus = True,
+ binary = "/fake/llama-server",
+ )
+ assert "--spec-draft-p-min" not in flags
+ assert backend.spec_draft_p_min is None
+
+
+def test_build_speculative_flags_p_min_emitted_via_auto_when_promoted(monkeypatch):
+ # Auto on an MTP GGUF auto-promotes to draft-mtp; p_min should still
+ # flow through because the resolved emission is MTP.
+ backend = _resolver_backend(monkeypatch)
+ flags = backend._build_speculative_flags(
+ speculative_type = "auto",
+ spec_draft_n_max = None,
+ spec_draft_p_min = 0.1,
+ extra_args = None,
+ model_identifier = _MTP_MODEL,
+ model_path = None,
+ gpus = True,
+ binary = "/fake/llama-server",
+ )
+ parsed = _flags_dict(flags)
+ assert parsed.get("--spec-type") == "draft-mtp"
+ assert parsed.get("--spec-draft-p-min") == "0.1000"
+
+
+# ── ngram-map-k / ngram-map-k4v ──────────────────────────────────
+
+
+@pytest.mark.parametrize("variant", ["ngram-map-k", "ngram-map-k4v"])
+def test_build_speculative_flags_ngram_map_variants(monkeypatch, variant):
+ backend = _resolver_backend(monkeypatch)
+ flags = backend._build_speculative_flags(
+ speculative_type = variant,
+ spec_draft_n_max = None,
+ extra_args = None,
+ model_identifier = _MTP_MODEL,
+ model_path = None,
+ gpus = True,
+ binary = "/fake/llama-server",
+ )
+ parsed = _flags_dict(flags)
+ assert parsed.get("--spec-type") == variant
+ assert parsed.get(f"--spec-{variant}-size-n") == "16"
+ assert parsed.get(f"--spec-{variant}-size-m") == "24"
+ assert parsed.get(f"--spec-{variant}-min-hits") == "1"
+ assert backend.requested_spec_mode == variant
+ assert backend.speculative_type == variant
+ # n-max/p-min are MTP-only knobs; ngram-map-* variants don't carry them.
+ assert "--spec-draft-n-max" not in flags
+ assert "--spec-draft-p-min" not in flags
+
+
+def test_build_speculative_flags_ngram_map_k_skipped_when_unsupported(monkeypatch):
+ backend = _resolver_backend(monkeypatch, supports_ngram_map_k = False)
+ flags = backend._build_speculative_flags(
+ speculative_type = "ngram-map-k",
+ spec_draft_n_max = None,
+ extra_args = None,
+ model_identifier = _MTP_MODEL,
+ model_path = None,
+ gpus = True,
+ binary = "/fake/llama-server",
+ )
+ assert "--spec-type" not in flags
+ assert backend.speculative_type is None
+ # Requested mode still round-trips so the UI / status can surface it.
+ assert backend.requested_spec_mode == "ngram-map-k"
+
+
+def test_canonicalize_spec_mode_recognises_ngram_map_variants():
+ assert _canonicalize_spec_mode("ngram-map-k") == "ngram-map-k"
+ assert _canonicalize_spec_mode("ngram-map-k4v") == "ngram-map-k4v"
+ assert _canonicalize_spec_mode("NGRAM-MAP-K4V") == "ngram-map-k4v"
diff --git a/studio/frontend/src/features/chat/chat-settings-sheet.tsx b/studio/frontend/src/features/chat/chat-settings-sheet.tsx
index 1ed0a88b63..7fe51d4d98 100644
--- a/studio/frontend/src/features/chat/chat-settings-sheet.tsx
+++ b/studio/frontend/src/features/chat/chat-settings-sheet.tsx
@@ -575,6 +575,11 @@ export function ChatSettingsPanel({
const loadedSpecDraftNMax = useChatRuntimeStore(
(s) => s.loadedSpecDraftNMax,
);
+ const specDraftPMin = useChatRuntimeStore((s) => s.specDraftPMin);
+ const setSpecDraftPMin = useChatRuntimeStore((s) => s.setSpecDraftPMin);
+ const loadedSpecDraftPMin = useChatRuntimeStore(
+ (s) => s.loadedSpecDraftPMin,
+ );
const modelRequiresTrustRemoteCode = useChatRuntimeStore(
(s) => s.modelRequiresTrustRemoteCode,
);
@@ -614,7 +619,9 @@ export function ChatSettingsPanel({
const ctxDirty = customContextLength !== null;
const specDirty = speculativeType !== loadedSpeculativeType;
const specDraftDirty = specDraftNMax !== loadedSpecDraftNMax;
- const modelSettingsDirty = kvDirty || ctxDirty || specDirty || specDraftDirty;
+ const specPMinDirty = specDraftPMin !== loadedSpecDraftPMin;
+ const modelSettingsDirty =
+ kvDirty || ctxDirty || specDirty || specDraftDirty || specPMinDirty;
const chatTemplateOverride = useChatRuntimeStore(
(s) => s.chatTemplateOverride,
);
@@ -1004,6 +1011,7 @@ export function ChatSettingsPanel({
setSpeculativeType(v);
if (v !== "mtp" && v !== "mtp+ngram") {
setSpecDraftNMax(null);
+ setSpecDraftPMin(null);
}
}}
>
@@ -1066,6 +1074,46 @@ export function ChatSettingsPanel({
/>
)}
+ {(speculativeType === "mtp" ||
+ speculativeType === "mtp+ngram") && (
+
+
+
+ Draft p-min
+
+
+ Min MTP draft probability (--spec-draft-p-min).
+ Drafts below this probability are rejected.
+ Range 0..1. Default: 0 (no filtering). Higher
+ values trade acceptance for lower wasted-draft
+ cost. Functional since llama.cpp #23269.
+
+