wangbo/ComfyUI
1
0
Fork 0
mirror of https://github.com/comfyanonymous/ComfyUI.git synced 2026-01-15 08:40:50 +08:00
Code Issues Actions 10 Packages Projects Releases Wiki Activity

Merge branch 'master' into master

Browse Source
This commit is contained in:
TheephopWS 2026-01-12 13:31:55 +08:00 committed by GitHub
commit af96939c95
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
10 changed files with 14 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
README.md
View File

@ -183,7 +183,7 @@ Simply download, extract with [7-Zip](https://7-zip.org) or with the windows exp
If you have trouble extracting it, right click the file -> properties -> unblock
Update your Nvidia drivers if it doesn't start.
The portable above currently comes with python 3.13 and pytorch cuda 13.0. Update your Nvidia drivers if it doesn't start.
#### Alternative Downloads:
@ -212,7 +212,7 @@ Python 3.14 works but you may encounter issues with the torch compile node. The
Python 3.13 is very well supported. If you have trouble with some custom node dependencies on 3.13 you can try 3.12
torch 2.4 and above is supported but some features might only work on newer versions. We generally recommend using the latest major version of pytorch unless it is less than 2 weeks old.
torch 2.4 and above is supported but some features might only work on newer versions. We generally recommend using the latest major version of pytorch with the latest cuda version unless it is less than 2 weeks old.
### Instructions:

2
comfy/model_detection.py
View File

@ -237,6 +237,8 @@ def detect_unet_config(state_dict, key_prefix, metadata=None):
else:
dit_config["vec_in_dim"] = None
dit_config["num_heads"] = dit_config["hidden_size"] // sum(dit_config["axes_dim"])
dit_config["depth"] = count_blocks(state_dict_keys, '{}double_blocks.'.format(key_prefix) + '{}.')
dit_config["depth_single_blocks"] = count_blocks(state_dict_keys, '{}single_blocks.'.format(key_prefix) + '{}.')
if '{}distilled_guidance_layer.0.norms.0.scale'.format(key_prefix) in state_dict_keys or '{}distilled_guidance_layer.norms.0.scale'.format(key_prefix) in state_dict_keys: #Chroma

2
comfy/model_management.py
View File

@ -368,7 +368,7 @@ try:
if any((a in arch) for a in ["gfx90a", "gfx942", "gfx1100", "gfx1101", "gfx1151"]): # TODO: more arches, TODO: gfx950
ENABLE_PYTORCH_ATTENTION = True
if rocm_version >= (7, 0):
if any((a in arch) for a in ["gfx1201"]):
if any((a in arch) for a in ["gfx1200", "gfx1201"]):
ENABLE_PYTORCH_ATTENTION = True
if torch_version_numeric >= (2, 7) and rocm_version >= (6, 4):
if any((a in arch) for a in ["gfx1200", "gfx1201", "gfx950"]): # TODO: more arches, "gfx942" gives error on pytorch nightly 2.10 1013 rocm7.0

4
comfy/sd.py
View File

@ -1059,9 +1059,9 @@ def detect_te_model(sd):
return TEModel.JINA_CLIP_2
if "encoder.block.23.layer.1.DenseReluDense.wi_1.weight" in sd:
weight = sd["encoder.block.23.layer.1.DenseReluDense.wi_1.weight"]
if weight.shape[-1] == 4096:
if weight.shape[0] == 10240:
return TEModel.T5_XXL
elif weight.shape[-1] == 2048:
elif weight.shape[0] == 5120:
return TEModel.T5_XL
if 'encoder.block.23.layer.1.DenseReluDense.wi.weight' in sd:
return TEModel.T5_XXL_OLD

2
comfy/text_encoders/cosmos.py
View File

@ -36,7 +36,7 @@ def te(dtype_t5=None, t5_quantization_metadata=None):
if t5_quantization_metadata is not None:
model_options = model_options.copy()
model_options["t5xxl_quantization_metadata"] = t5_quantization_metadata
if dtype is None:
if dtype_t5 is not None:
dtype = dtype_t5
super().__init__(device=device, dtype=dtype, model_options=model_options)
return CosmosTEModel_

2
comfy/text_encoders/genmo.py
View File

@ -32,7 +32,7 @@ def mochi_te(dtype_t5=None, t5_quantization_metadata=None):
if t5_quantization_metadata is not None:
model_options = model_options.copy()
model_options["t5xxl_quantization_metadata"] = t5_quantization_metadata
if dtype is None:
if dtype_t5 is not None:
dtype = dtype_t5
super().__init__(device=device, dtype=dtype, model_options=model_options)
return MochiTEModel_

2
comfy/text_encoders/pixart_t5.py
View File

@ -36,7 +36,7 @@ def pixart_te(dtype_t5=None, t5_quantization_metadata=None):
if t5_quantization_metadata is not None:
model_options = model_options.copy()
model_options["t5xxl_quantization_metadata"] = t5_quantization_metadata
if dtype is None:
if dtype_t5 is not None:
dtype = dtype_t5
super().__init__(device=device, dtype=dtype, model_options=model_options)
return PixArtTEModel_

2
comfy_api_nodes/nodes_kling.py
View File

@ -567,7 +567,7 @@ async def execute_lipsync(
# Upload the audio file to Comfy API and get download URL
if audio:
audio_url = await upload_audio_to_comfyapi(
cls, audio, container_format="mp3", codec_name="libmp3lame", mime_type="audio/mpeg", filename="output.mp3"
cls, audio, container_format="mp3", codec_name="libmp3lame", mime_type="audio/mpeg"
)
logging.info("Uploaded audio to Comfy API. URL: %s", audio_url)
else:

4
comfy_api_nodes/util/conversions.py
View File

@ -55,7 +55,7 @@ def image_tensor_pair_to_batch(image1: torch.Tensor, image2: torch.Tensor) -> to
def tensor_to_bytesio(
image: torch.Tensor,
name: str | None = None,
*,
total_pixels: int = 2048 * 2048,
mime_type: str = "image/png",
) -> BytesIO:
@ -75,7 +75,7 @@ def tensor_to_bytesio(
pil_image = tensor_to_pil(image, total_pixels=total_pixels)
img_binary = pil_to_bytesio(pil_image, mime_type=mime_type)
img_binary.name = f"{name if name else uuid.uuid4()}.{mimetype_to_extension(mime_type)}"
img_binary.name = f"{uuid.uuid4()}.{mimetype_to_extension(mime_type)}"
return img_binary

3
comfy_api_nodes/util/upload_helpers.py
View File

@ -82,7 +82,6 @@ async def upload_audio_to_comfyapi(
container_format: str = "mp4",
codec_name: str = "aac",
mime_type: str = "audio/mp4",
filename: str = "uploaded_audio.mp4",
) -> str:
"""
Uploads a single audio input to ComfyUI API and returns its download URL.
@ -92,7 +91,7 @@ async def upload_audio_to_comfyapi(
waveform: torch.Tensor = audio["waveform"]
audio_data_np = audio_tensor_to_contiguous_ndarray(waveform)
audio_bytes_io = audio_ndarray_to_bytesio(audio_data_np, sample_rate, container_format, codec_name)
return await upload_file_to_comfyapi(cls, audio_bytes_io, filename, mime_type)
return await upload_file_to_comfyapi(cls, audio_bytes_io, f"{uuid.uuid4()}.{container_format}", mime_type)
async def upload_video_to_comfyapi(