ceos/unsloth
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

[pre-commit.ci] auto fixes from pre-commit.com hooks

Browse source 
for more information, see https://pre-commit.ci
This commit is contained in:
pre-commit-ci[bot] 2026-06-28 04:31:35 +00:00
commit 86b3a5d2d1
9 changed files with 82 additions and 36 deletions
Download patch file Download diff file Expand all files Collapse all files

1
studio/backend/core/inference/diffusion_engine_router.py
View file

@ -56,7 +56,6 @@ def get_active_diffusion_engine() -> Any:
"""The engine object the active selection points at (defaults to diffusers)."""
if _active_engine_name == ENGINE_SD_CPP:
from core.inference.sd_cpp_backend import get_sd_cpp_backend
return get_sd_cpp_backend()
from core.inference.diffusion import get_diffusion_backend

6
studio/backend/core/inference/diffusion_families.py
View file

@ -108,7 +108,11 @@ _FAMILIES: tuple[DiffusionFamily, ...] = (
# The Qwen2.5-VL text encoder as a Q4_K_M GGUF keeps the CPU RAM win (the
# bf16 safetensors encoder is ~15 GB). sd-cli's --qwen2vl is an alias of --llm.
sd_cpp_text_encoders = (
("unsloth/Qwen2.5-VL-7B-Instruct-GGUF", "Qwen2.5-VL-7B-Instruct-Q4_K_M.gguf", "qwen2vl"),
(
"unsloth/Qwen2.5-VL-7B-Instruct-GGUF",
"Qwen2.5-VL-7B-Instruct-Q4_K_M.gguf",
"qwen2vl",
),
),
),
DiffusionFamily(

47
studio/backend/core/inference/sd_cpp_backend.py
View file

@ -141,7 +141,9 @@ def _estimate_eta(total_steps: int, step: int, first_step_at: float, now: float)
return max(0.0, (total_steps - step) * per_step)
def _map_guidance(fam: DiffusionFamily, guidance: Optional[float]) -> tuple[Optional[float], Optional[float]]:
def _map_guidance(
fam: DiffusionFamily, guidance: Optional[float]
) -> tuple[Optional[float], Optional[float]]:
"""(cfg_scale, guidance) for sd-cli from the single diffusers ``guidance`` value.
FLUX families take a distilled embedded ``--guidance``; everyone else uses real
@ -309,12 +311,13 @@ class SdCppDiffusionBackend:
specs.append((terepo, tefile, kind))
return specs
def _set_expected_bytes(self, assets: list[tuple[str, str, str]], hf_token: Optional[str]) -> None:
def _set_expected_bytes(
self, assets: list[tuple[str, str, str]], hf_token: Optional[str]
) -> None:
"""Best-effort total download size for the progress bar (0 if unknown)."""
total = 0
try:
from huggingface_hub import HfApi
api = HfApi(token = hf_token)
for repo, fn, _ in assets:
if Path(repo).expanduser().exists():
@ -402,9 +405,7 @@ class SdCppDiffusionBackend:
else:
seed = int(seed)
cfg_scale, flux_guidance = _map_guidance(state.family, guidance)
vae_extra = (
["--vae-format", state.vae_format] if state.vae_format else None
)
vae_extra = ["--vae-format", state.vae_format] if state.vae_format else None
self._gen = _SdGen(total_steps = int(steps))
images = []
@ -461,14 +462,18 @@ class SdCppDiffusionBackend:
gen.step = min(int(a), gen.total_steps)
if gen.first_step_at == 0.0:
gen.first_step_at = now
gen.eta_seconds = _estimate_eta(
gen.total_steps, gen.step, gen.first_step_at, now
)
gen.eta_seconds = _estimate_eta(gen.total_steps, gen.step, gen.first_step_at, now)
def generate_progress(self) -> dict[str, Any]:
gen = self._gen
if gen is None or gen.total_steps <= 0:
return {"active": False, "step": 0, "total_steps": 0, "fraction": 0.0, "eta_seconds": None}
return {
"active": False,
"step": 0,
"total_steps": 0,
"fraction": 0.0,
"eta_seconds": None,
}
return {
"active": True,
"step": gen.step,
@ -493,11 +498,23 @@ class SdCppDiffusionBackend:
state = self._state
if state is None:
return {
"loaded": False, "repo_id": None, "family": None, "base_repo": None,
"device": None, "dtype": None, "cpu_offload": False, "offload_policy": None,
"vae_tiling": False, "memory_mode": None, "speed_mode": None, "speed_optims": [],
"text_encoder_quant": None, "transformer_quant": None, "attention_backend": None,
"transformer_cache": None, "engine": "sd_cpp",
"loaded": False,
"repo_id": None,
"family": None,
"base_repo": None,
"device": None,
"dtype": None,
"cpu_offload": False,
"offload_policy": None,
"vae_tiling": False,
"memory_mode": None,
"speed_mode": None,
"speed_optims": [],
"text_encoder_quant": None,
"transformer_quant": None,
"attention_backend": None,
"transformer_cache": None,
"engine": "sd_cpp",
}
return {
"loaded": True,

12
studio/backend/core/inference/sd_cpp_engine.py
View file

@ -250,7 +250,11 @@ class SdCppEngine:
extra_args = merged_extra,
)
return self._run(
cmd, output_path, timeout = timeout, env = env, on_log = on_log,
cmd,
output_path,
timeout = timeout,
env = env,
on_log = on_log,
cancel_event = cancel_event,
)
@ -275,7 +279,11 @@ class SdCppEngine:
extra_args = extra_args,
)
return self._run(
cmd, output_path, timeout = timeout, env = env, on_log = on_log,
cmd,
output_path,
timeout = timeout,
env = env,
on_log = on_log,
cancel_event = cancel_event,
)

4
studio/backend/models/inference.py
View file

@ -1904,9 +1904,7 @@ class DiffusionStatusResponse(BaseModel):
"_native_cudnn), or null for the default SDPA",
)
transformer_cache: Optional[str] = Field(None, description = "Step cache engaged: fbcache | null")
engine: Optional[str] = Field(
None, description = "Active diffusion engine: diffusers | sd_cpp"
)
engine: Optional[str] = Field(None, description = "Active diffusion engine: diffusers | sd_cpp")
fallback_reason: Optional[str] = Field(
None,
description = "Why diffusers was chosen over the native sd.cpp engine (null when none)",

4
studio/backend/routes/inference.py
View file

@ -10076,9 +10076,7 @@ async def load_diffusion_model(
# Pick the engine for this host (diffusers on GPU, native sd.cpp with no GPU),
# installing the sd-cli binary if needed -- all BEFORE evicting chat, so a
# native fallback never strands a half-loaded state.
engine = await asyncio.to_thread(
select_and_activate_engine, fam, hf_token = request.hf_token
)
engine = await asyncio.to_thread(select_and_activate_engine, fam, hf_token = request.hf_token)
# Now take the GPU from the chat backend, then kick the (slow) load onto a
# background thread and return at once — the client polls images/load-progress.
await asyncio.to_thread(acquire_for, DIFFUSION)

6
studio/backend/tests/test_diffusion_engine_router.py
View file

@ -28,7 +28,8 @@ def _clean_env_and_state(monkeypatch):
# A light status-capable stub so neither selection nor active_status() imports the
# heavy diffusers/sd.cpp backends; the active engine NAME comes from module state.
monkeypatch.setattr(
r, "get_active_diffusion_engine",
r,
"get_active_diffusion_engine",
lambda: SimpleNamespace(status = lambda: {"loaded": False, "repo_id": None}),
)
yield
@ -36,7 +37,8 @@ def _clean_env_and_state(monkeypatch):
def _set_device(monkeypatch, backend):
monkeypatch.setattr(
r, "resolve_diffusion_device_target",
r,
"resolve_diffusion_device_target",
lambda: SimpleNamespace(backend = backend, device = backend),
)

15
studio/backend/tests/test_diffusion_routes.py
View file

@ -123,11 +123,13 @@ def client(monkeypatch, tmp_path):
# Delegate to whatever get_diffusion_backend currently returns, so per-test
# re-patches of the backend still flow through the routes.
monkeypatch.setattr(
engine_router, "select_and_activate_engine",
engine_router,
"select_and_activate_engine",
lambda fam, **kw: diffusion_module.get_diffusion_backend(),
)
monkeypatch.setattr(
engine_router, "get_active_diffusion_engine",
engine_router,
"get_active_diffusion_engine",
lambda: diffusion_module.get_diffusion_backend(),
)
monkeypatch.setattr(engine_router, "_active_engine_name", "diffusers")
@ -450,8 +452,10 @@ def test_load_routes_to_sd_cpp_on_cpu(monkeypatch, tmp_path):
import core.inference.sd_cpp_backend as sd_backend
for e in (
"UNSLOTH_DIFFUSION_ENGINE", "UNSLOTH_DIFFUSION_SD_CPP",
"UNSLOTH_DIFFUSION_SD_CPP_MPS", "UNSLOTH_DIFFUSION_SD_CPP_INSTALL",
"UNSLOTH_DIFFUSION_ENGINE",
"UNSLOTH_DIFFUSION_SD_CPP",
"UNSLOTH_DIFFUSION_SD_CPP_MPS",
"UNSLOTH_DIFFUSION_SD_CPP_INSTALL",
):
monkeypatch.delenv(e, raising = False)
@ -459,7 +463,8 @@ def test_load_routes_to_sd_cpp_on_cpu(monkeypatch, tmp_path):
monkeypatch.setattr(diffusion_module, "get_diffusion_backend", lambda: validator)
# Force the router's decision inputs: CPU device + an available binary.
monkeypatch.setattr(
engine_router, "resolve_diffusion_device_target",
engine_router,
"resolve_diffusion_device_target",
lambda: SimpleNamespace(backend = "cpu", device = "cpu"),
)
monkeypatch.setattr(engine_router, "ensure_sd_cpp_binary", lambda **_: "/x/sd-cli")

23
studio/backend/tests/test_sd_cpp_backend.py
View file

@ -24,7 +24,12 @@ from core.inference.sd_cpp_engine import SdCppCancelled
class _FakeEngine:
"""Stands in for SdCppEngine: writes a 1x1 PNG and records the args."""
def __init__(self, *, fail=None, cancel_on_call=False):
def __init__(
self,
*,
fail = None,
cancel_on_call = False,
):
self.calls = []
self.fail = fail
self.cancel_on_call = cancel_on_call
@ -35,7 +40,15 @@ class _FakeEngine:
def version(self, **_):
return "fake sd-cli"
def generate(self, files, params, *, output_path, cancel_event=None, **kw):
def generate(
self,
files,
params,
*,
output_path,
cancel_event = None,
**kw,
):
self.calls.append((files, params, output_path, kw))
if self.cancel_on_call and cancel_event is not None:
cancel_event.set()
@ -49,7 +62,7 @@ class _FakeEngine:
return Path(output_path)
def _loaded_backend(fam_name="z-image", engine=None):
def _loaded_backend(fam_name = "z-image", engine = None):
b = SdCppDiffusionBackend(engine = engine or _FakeEngine())
fam = detect_family(fam_name)
b._state = bk._SdState(
@ -57,7 +70,9 @@ def _loaded_backend(fam_name="z-image", engine=None):
base_repo = fam.base_repo,
family = fam,
device = "cpu",
files = SdCppModelFiles(diffusion_model = "/m/z.gguf", vae = "/m/vae.safetensors", llm = "/m/llm.safetensors"),
files = SdCppModelFiles(
diffusion_model = "/m/z.gguf", vae = "/m/vae.safetensors", llm = "/m/llm.safetensors"
),
vae_format = fam.sd_cpp_vae_format,
)
return b