mirror of
https://github.com/comfyanonymous/ComfyUI.git
synced 2026-01-10 06:10:50 +08:00
Merge branch 'master' of github.com:comfyanonymous/ComfyUI
This commit is contained in:
doctorpangloss
commit
7ba86929f7
@ -1,3 +1,3 @@
|
||||
# This file is automatically generated by the build process when version is
|
||||
# updated in pyproject.toml.
|
||||
__version__ = "0.3.47"
|
||||
__version__ = "0.3.48"
|
||||
|
||||
@ -131,10 +131,21 @@ class ModelFileManager:
|
||||
|
||||
for file_name in filenames:
|
||||
try:
|
||||
relative_path = os.path.relpath(os.path.join(dirpath, file_name), directory)
|
||||
result.append(relative_path)
|
||||
except:
|
||||
logging.warning(f"Warning: Unable to access {file_name}. Skipping this file.")
|
||||
full_path = os.path.join(dirpath, file_name)
|
||||
relative_path = os.path.relpath(full_path, directory)
|
||||
|
||||
# Get file metadata
|
||||
file_info = {
|
||||
"name": relative_path,
|
||||
"pathIndex": pathIndex,
|
||||
"modified": os.path.getmtime(full_path), # Add modification time
|
||||
"created": os.path.getctime(full_path), # Add creation time
|
||||
"size": os.path.getsize(full_path) # Add file size
|
||||
}
|
||||
result.append(file_info)
|
||||
|
||||
except Exception as e:
|
||||
logging.warning(f"Warning: Unable to access {file_name}. Error: {e}. Skipping this file.")
|
||||
continue
|
||||
|
||||
for d in subdirs:
|
||||
@ -145,7 +156,7 @@ class ModelFileManager:
|
||||
logging.warning(f"Warning: Unable to access {path}. Skipping this path.")
|
||||
continue
|
||||
|
||||
return [{"name": f, "pathIndex": pathIndex} for f in result], dirs, time.perf_counter()
|
||||
return result, dirs, time.perf_counter()
|
||||
|
||||
def get_model_previews(self, filepath: str) -> list[str | BytesIO]:
|
||||
dirname = os.path.dirname(filepath)
|
||||
|
||||
@ -23,13 +23,15 @@ class FileInfo(TypedDict):
|
||||
path: str
|
||||
size: int
|
||||
modified: float
|
||||
created: int
|
||||
|
||||
|
||||
def get_file_info(path: str, relative_to: str) -> FileInfo:
|
||||
return {
|
||||
"path": os.path.relpath(path, relative_to).replace(os.sep, '/'),
|
||||
"size": os.path.getsize(path),
|
||||
"modified": os.path.getmtime(path)
|
||||
"modified": os.path.getmtime(path),
|
||||
"created": os.path.getctime(path)
|
||||
}
|
||||
|
||||
|
||||
|
||||
@ -3,6 +3,7 @@ import math
|
||||
import torch
|
||||
|
||||
from . import utils
|
||||
import logging
|
||||
|
||||
|
||||
class CONDRegular:
|
||||
@ -12,12 +13,15 @@ class CONDRegular:
|
||||
def _copy_with(self, cond):
|
||||
return self.__class__(cond)
|
||||
|
||||
def process_cond(self, batch_size, device, **kwargs):
|
||||
return self._copy_with(utils.repeat_to_batch_size(self.cond, batch_size).to(device))
|
||||
def process_cond(self, batch_size, **kwargs):
|
||||
return self._copy_with(utils.repeat_to_batch_size(self.cond, batch_size))
|
||||
|
||||
def can_concat(self, other):
|
||||
if self.cond.shape != other.cond.shape:
|
||||
return False
|
||||
if self.cond.device != other.cond.device:
|
||||
logging.warning("WARNING: conds not on same device, skipping concat.")
|
||||
return False
|
||||
return True
|
||||
|
||||
def concat(self, others):
|
||||
@ -31,14 +35,14 @@ class CONDRegular:
|
||||
|
||||
|
||||
class CONDNoiseShape(CONDRegular):
|
||||
def process_cond(self, batch_size, device, area, **kwargs):
|
||||
def process_cond(self, batch_size, area, **kwargs):
|
||||
data = self.cond
|
||||
if area is not None:
|
||||
dims = len(area) // 2
|
||||
for i in range(dims):
|
||||
data = data.narrow(i + 2, area[i + dims], area[i])
|
||||
|
||||
return self._copy_with(utils.repeat_to_batch_size(data, batch_size).to(device))
|
||||
return self._copy_with(utils.repeat_to_batch_size(data, batch_size))
|
||||
|
||||
|
||||
class CONDCrossAttn(CONDRegular):
|
||||
@ -53,6 +57,9 @@ class CONDCrossAttn(CONDRegular):
|
||||
diff = mult_min // min(s1[1], s2[1])
|
||||
if diff > 4: # arbitrary limit on the padding because it's probably going to impact performance negatively if it's too much
|
||||
return False
|
||||
if self.cond.device != other.cond.device:
|
||||
logging.warning("WARNING: conds not on same device: skipping concat.")
|
||||
return False
|
||||
return True
|
||||
|
||||
def concat(self, others):
|
||||
@ -75,7 +82,7 @@ class CONDConstant(CONDRegular):
|
||||
def __init__(self, cond):
|
||||
self.cond = cond
|
||||
|
||||
def process_cond(self, batch_size, device, **kwargs):
|
||||
def process_cond(self, batch_size, **kwargs):
|
||||
return self._copy_with(self.cond)
|
||||
|
||||
def can_concat(self, other):
|
||||
@ -94,10 +101,10 @@ class CONDList(CONDRegular):
|
||||
def __init__(self, cond):
|
||||
self.cond = cond
|
||||
|
||||
def process_cond(self, batch_size, device, **kwargs):
|
||||
def process_cond(self, batch_size, **kwargs):
|
||||
out = []
|
||||
for c in self.cond:
|
||||
out.append(utils.repeat_to_batch_size(c, batch_size).to(device))
|
||||
out.append(utils.repeat_to_batch_size(c, batch_size))
|
||||
|
||||
return self._copy_with(out)
|
||||
|
||||
|
||||
@ -36,6 +36,8 @@ from .ldm.cascade import controlnet as cascade_controlnet
|
||||
from .ldm.flux import controlnet as controlnet_flux
|
||||
from .ldm.hydit.controlnet import HunYuanControlNet
|
||||
from .t2i_adapter import adapter
|
||||
from .model_base import convert_tensor
|
||||
from .model_management import cast_to_device
|
||||
|
||||
if TYPE_CHECKING:
|
||||
from .hooks import HookGroup
|
||||
@ -43,7 +45,6 @@ if TYPE_CHECKING:
|
||||
|
||||
def broadcast_image_to(tensor, target_batch_size, batched_number):
|
||||
current_batch_size = tensor.shape[0]
|
||||
# print(current_batch_size, target_batch_size)
|
||||
if current_batch_size == 1:
|
||||
return tensor
|
||||
|
||||
@ -268,12 +269,12 @@ class ControlNet(ControlBase):
|
||||
for c in self.extra_conds:
|
||||
temp = cond.get(c, None)
|
||||
if temp is not None:
|
||||
extra[c] = temp.to(dtype)
|
||||
extra[c] = convert_tensor(temp, dtype, x_noisy.device)
|
||||
|
||||
timestep = self.model_sampling_current.timestep(t)
|
||||
x_noisy = self.model_sampling_current.calculate_input(t, x_noisy)
|
||||
|
||||
control = self.control_model(x=x_noisy.to(dtype), hint=self.cond_hint, timesteps=timestep.to(dtype), context=context.to(dtype), **extra)
|
||||
control = self.control_model(x=x_noisy.to(dtype), hint=self.cond_hint, timesteps=timestep.to(dtype), context=cast_to_device(context, x_noisy.device, dtype), **extra)
|
||||
return self.control_merge(control, control_prev, output_dtype=None)
|
||||
|
||||
def copy(self):
|
||||
|
||||
@ -122,10 +122,12 @@ class ComfyUIModel(Protocol):
|
||||
...
|
||||
|
||||
|
||||
def convert_tensor(extra, dtype):
|
||||
def convert_tensor(extra, dtype, device):
|
||||
if hasattr(extra, "dtype"):
|
||||
if extra.dtype != torch.int and extra.dtype != torch.long:
|
||||
extra = extra.to(dtype)
|
||||
extra = model_management.cast_to_device(extra, device, dtype)
|
||||
else:
|
||||
extra = model_management.cast_to_device(extra, device, None)
|
||||
return extra
|
||||
|
||||
|
||||
@ -181,7 +183,7 @@ class BaseModel(torch.nn.Module):
|
||||
xc = self.model_sampling.calculate_input(sigma, x)
|
||||
|
||||
if c_concat is not None:
|
||||
xc = torch.cat([xc] + [c_concat], dim=1)
|
||||
xc = torch.cat([xc] + [model_management.cast_to_device(c_concat, xc.device, xc.dtype)], dim=1)
|
||||
|
||||
context = c_crossattn
|
||||
dtype = self.get_dtype()
|
||||
@ -190,20 +192,21 @@ class BaseModel(torch.nn.Module):
|
||||
dtype = self.manual_cast_dtype
|
||||
|
||||
xc = xc.to(dtype)
|
||||
device = xc.device
|
||||
t = self.model_sampling.timestep(t).float()
|
||||
if context is not None:
|
||||
context = context.to(dtype)
|
||||
context = model_management.cast_to_device(context, device, dtype)
|
||||
|
||||
extra_conds = {}
|
||||
for o in kwargs:
|
||||
extra = kwargs[o]
|
||||
|
||||
if hasattr(extra, "dtype"):
|
||||
extra = convert_tensor(extra, dtype)
|
||||
extra = convert_tensor(extra, dtype, device)
|
||||
elif isinstance(extra, list):
|
||||
ex = []
|
||||
for ext in extra:
|
||||
ex.append(convert_tensor(ext, dtype))
|
||||
ex.append(convert_tensor(ext, dtype, device))
|
||||
extra = ex
|
||||
extra_conds[o] = extra
|
||||
|
||||
@ -423,7 +426,7 @@ class SD21UNCLIP(BaseModel):
|
||||
unclip_conditioning = kwargs.get("unclip_conditioning", None)
|
||||
device = kwargs["device"]
|
||||
if unclip_conditioning is None:
|
||||
return torch.zeros((1, self.adm_channels))
|
||||
return torch.zeros((1, self.adm_channels), device=device)
|
||||
else:
|
||||
return unclip_adm(unclip_conditioning, device, self.noise_augmentor, kwargs.get("unclip_noise_augment_merge", 0.05), kwargs.get("seed", 0) - 10)
|
||||
|
||||
@ -647,9 +650,11 @@ class IP2P(BaseModel):
|
||||
|
||||
if image is None:
|
||||
image = torch.zeros_like(noise)
|
||||
else:
|
||||
image = image.to(device=device)
|
||||
|
||||
if image.shape[1:] != noise.shape[1:]:
|
||||
image = utils.common_upscale(image.to(device), noise.shape[-1], noise.shape[-2], "bilinear", "center")
|
||||
image = utils.common_upscale(image, noise.shape[-1], noise.shape[-2], "bilinear", "center")
|
||||
|
||||
image = utils.resize_to_batch_size(image, noise.shape[0])
|
||||
return self.process_ip2p_image_in(image)
|
||||
@ -730,7 +735,7 @@ class StableCascade_B(BaseModel):
|
||||
# size of prior doesn't really matter if zeros because it gets resized but I still want it to get batched
|
||||
prior = kwargs.get("stable_cascade_prior", torch.zeros((1, 16, (noise.shape[2] * 4) // 42, (noise.shape[3] * 4) // 42), dtype=noise.dtype, layout=noise.layout, device=noise.device))
|
||||
|
||||
out["effnet"] = conds.CONDRegular(prior)
|
||||
out["effnet"] = conds.CONDRegular(prior.to(device=noise.device))
|
||||
out["sca"] = conds.CONDRegular(torch.zeros((1,)))
|
||||
return out
|
||||
|
||||
@ -1203,10 +1208,10 @@ class WAN21_Vace(WAN21):
|
||||
|
||||
vace_frames_out = []
|
||||
for j in range(len(vace_frames)):
|
||||
vf = vace_frames[j].clone()
|
||||
vf = vace_frames[j].to(device=noise.device, dtype=noise.dtype, copy=True)
|
||||
for i in range(0, vf.shape[1], 16):
|
||||
vf[:, i:i + 16] = self.process_latent_in(vf[:, i:i + 16])
|
||||
vf = torch.cat([vf, mask[j]], dim=1)
|
||||
vf = torch.cat([vf, mask[j].to(device=noise.device, dtype=noise.dtype)], dim=1)
|
||||
vace_frames_out.append(vf)
|
||||
|
||||
vace_frames = torch.stack(vace_frames_out, dim=1)
|
||||
|
||||
@ -96,7 +96,7 @@ def get_area_and_mult(conds, x_in, timestep_in):
|
||||
conditioning = {}
|
||||
model_conds = conds["model_conds"]
|
||||
for c in model_conds:
|
||||
conditioning[c] = model_conds[c].process_cond(batch_size=x_in.shape[0], device=x_in.device, area=area)
|
||||
conditioning[c] = model_conds[c].process_cond(batch_size=x_in.shape[0], area=area)
|
||||
|
||||
hooks = conds.get('hooks', None)
|
||||
control = conds.get('control', None)
|
||||
|
||||
@ -1,6 +1,6 @@
|
||||
[project]
|
||||
name = "comfyui"
|
||||
version = "0.3.47"
|
||||
version = "0.3.48"
|
||||
description = "An installable version of ComfyUI"
|
||||
readme = "README.md"
|
||||
authors = [
|
||||
@ -19,7 +19,7 @@ classifiers = [
|
||||
|
||||
dependencies = [
|
||||
"comfyui-frontend-package>=1.23.4",
|
||||
"comfyui-workflow-templates>=0.1.41",
|
||||
"comfyui-workflow-templates>=0.1.47",
|
||||
"comfyui-embedded-docs>=0.2.4",
|
||||
"torch",
|
||||
"torchvision",
|
||||
|
||||
@ -1,10 +1,9 @@
|
||||
import os
|
||||
import tempfile
|
||||
|
||||
import pytest
|
||||
|
||||
import tempfile
|
||||
from unittest.mock import patch
|
||||
|
||||
|
||||
@pytest.fixture(scope="module")
|
||||
def file_extensions():
|
||||
return {
|
||||
@ -35,11 +34,11 @@ def patched_mimetype_cache(file_extensions):
|
||||
v: k for (k, ext) in file_extensions.items() for v in ext
|
||||
})
|
||||
|
||||
|
||||
with patch("comfy.component_model.folder_path_types.extension_mimetypes_cache", new_cache):
|
||||
yield
|
||||
|
||||
|
||||
@pytest.mark.skip("idiosyncratic")
|
||||
def test_categorizes_all_correctly(mock_dir, file_extensions, patched_mimetype_cache):
|
||||
from comfy.cmd.folder_paths import filter_files_content_types
|
||||
files = os.listdir(mock_dir)
|
||||
@ -49,6 +48,7 @@ def test_categorizes_all_correctly(mock_dir, file_extensions, patched_mimetype_c
|
||||
assert f"sample_{content_type}.{extension}" in filtered_files
|
||||
|
||||
|
||||
@pytest.mark.skip("idiosyncratic")
|
||||
def test_categorizes_all_uniquely(mock_dir, file_extensions, patched_mimetype_cache):
|
||||
from comfy.cmd.folder_paths import filter_files_content_types
|
||||
|
||||
|
||||
Loading…
Reference in New Issue
Block a user