From c4a8cf60ab5d6eaf052b7a08f5ee97104acf7a2f Mon Sep 17 00:00:00 2001 From: AustinMroz Date: Tue, 30 Sep 2025 22:12:32 -0700 Subject: [PATCH 01/37] Bump frontend to 1.27.7 (#10133) --- requirements.txt | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/requirements.txt b/requirements.txt index 45d3e1607164..588c5dcf01cc 100644 --- a/requirements.txt +++ b/requirements.txt @@ -1,4 +1,4 @@ -comfyui-frontend-package==1.26.13 +comfyui-frontend-package==1.27.7 comfyui-workflow-templates==0.1.91 comfyui-embedded-docs==0.2.6 torch From 638097829d2352a1c78ab4fbb1e028d1e7cff012 Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Wed, 1 Oct 2025 09:00:22 +0300 Subject: [PATCH 02/37] convert nodes_audio_encoder.py to V3 schema (#10123) --- comfy_api/latest/_io.py | 1 + comfy_extras/nodes_audio_encoder.py | 68 ++++++++++++++++++----------- 2 files changed, 44 insertions(+), 25 deletions(-) diff --git a/comfy_api/latest/_io.py b/comfy_api/latest/_io.py index 4826818df860..2d95cffd6dc5 100644 --- a/comfy_api/latest/_io.py +++ b/comfy_api/latest/_io.py @@ -1605,6 +1605,7 @@ class _IO: Model = Model ClipVision = ClipVision ClipVisionOutput = ClipVisionOutput + AudioEncoder = AudioEncoder AudioEncoderOutput = AudioEncoderOutput StyleModel = StyleModel Gligen = Gligen diff --git a/comfy_extras/nodes_audio_encoder.py b/comfy_extras/nodes_audio_encoder.py index 39a140fef182..13aacd41a47e 100644 --- a/comfy_extras/nodes_audio_encoder.py +++ b/comfy_extras/nodes_audio_encoder.py @@ -1,44 +1,62 @@ import folder_paths import comfy.audio_encoders.audio_encoders import comfy.utils +from typing_extensions import override +from comfy_api.latest import ComfyExtension, io -class AudioEncoderLoader: +class AudioEncoderLoader(io.ComfyNode): @classmethod - def INPUT_TYPES(s): - return {"required": { "audio_encoder_name": (folder_paths.get_filename_list("audio_encoders"), ), - }} - RETURN_TYPES = ("AUDIO_ENCODER",) - FUNCTION = "load_model" + def define_schema(cls) -> io.Schema: + return io.Schema( + node_id="AudioEncoderLoader", + category="loaders", + inputs=[ + io.Combo.Input( + "audio_encoder_name", + options=folder_paths.get_filename_list("audio_encoders"), + ), + ], + outputs=[io.AudioEncoder.Output()], + ) - CATEGORY = "loaders" - - def load_model(self, audio_encoder_name): + @classmethod + def execute(cls, audio_encoder_name) -> io.NodeOutput: audio_encoder_name = folder_paths.get_full_path_or_raise("audio_encoders", audio_encoder_name) sd = comfy.utils.load_torch_file(audio_encoder_name, safe_load=True) audio_encoder = comfy.audio_encoders.audio_encoders.load_audio_encoder_from_sd(sd) if audio_encoder is None: raise RuntimeError("ERROR: audio encoder file is invalid and does not contain a valid model.") - return (audio_encoder,) + return io.NodeOutput(audio_encoder) -class AudioEncoderEncode: +class AudioEncoderEncode(io.ComfyNode): @classmethod - def INPUT_TYPES(s): - return {"required": { "audio_encoder": ("AUDIO_ENCODER",), - "audio": ("AUDIO",), - }} - RETURN_TYPES = ("AUDIO_ENCODER_OUTPUT",) - FUNCTION = "encode" - - CATEGORY = "conditioning" + def define_schema(cls) -> io.Schema: + return io.Schema( + node_id="AudioEncoderEncode", + category="conditioning", + inputs=[ + io.AudioEncoder.Input("audio_encoder"), + io.Audio.Input("audio"), + ], + outputs=[io.AudioEncoderOutput.Output()], + ) - def encode(self, audio_encoder, audio): + @classmethod + def execute(cls, audio_encoder, audio) -> io.NodeOutput: output = audio_encoder.encode_audio(audio["waveform"], audio["sample_rate"]) - return (output,) + return io.NodeOutput(output) + + +class AudioEncoder(ComfyExtension): + @override + async def get_node_list(self) -> list[type[io.ComfyNode]]: + return [ + AudioEncoderLoader, + AudioEncoderEncode, + ] -NODE_CLASS_MAPPINGS = { - "AudioEncoderLoader": AudioEncoderLoader, - "AudioEncoderEncode": AudioEncoderEncode, -} +async def comfy_entrypoint() -> AudioEncoder: + return AudioEncoder() From 7eb7160db487feb891ceabdf985b09f9a8091869 Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Wed, 1 Oct 2025 22:16:59 +0300 Subject: [PATCH 03/37] convert nodes_gits.py to V3 schema (#9949) --- comfy_extras/nodes_gits.py | 49 ++++++++++++++++++++++++-------------- 1 file changed, 31 insertions(+), 18 deletions(-) diff --git a/comfy_extras/nodes_gits.py b/comfy_extras/nodes_gits.py index 47b1dd049702..25367560a605 100644 --- a/comfy_extras/nodes_gits.py +++ b/comfy_extras/nodes_gits.py @@ -1,6 +1,8 @@ # from https://github.com/zju-pi/diff-sampler/tree/main/gits-main import numpy as np import torch +from typing_extensions import override +from comfy_api.latest import ComfyExtension, io def loglinear_interp(t_steps, num_steps): """ @@ -333,25 +335,28 @@ def loglinear_interp(t_steps, num_steps): ], } -class GITSScheduler: +class GITSScheduler(io.ComfyNode): @classmethod - def INPUT_TYPES(s): - return {"required": - {"coeff": ("FLOAT", {"default": 1.20, "min": 0.80, "max": 1.50, "step": 0.05}), - "steps": ("INT", {"default": 10, "min": 2, "max": 1000}), - "denoise": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 1.0, "step": 0.01}), - } - } - RETURN_TYPES = ("SIGMAS",) - CATEGORY = "sampling/custom_sampling/schedulers" + def define_schema(cls): + return io.Schema( + node_id="GITSScheduler", + category="sampling/custom_sampling/schedulers", + inputs=[ + io.Float.Input("coeff", default=1.20, min=0.80, max=1.50, step=0.05), + io.Int.Input("steps", default=10, min=2, max=1000), + io.Float.Input("denoise", default=1.0, min=0.0, max=1.0, step=0.01), + ], + outputs=[ + io.Sigmas.Output(), + ], + ) - FUNCTION = "get_sigmas" - - def get_sigmas(self, coeff, steps, denoise): + @classmethod + def execute(cls, coeff, steps, denoise): total_steps = steps if denoise < 1.0: if denoise <= 0.0: - return (torch.FloatTensor([]),) + return io.NodeOutput(torch.FloatTensor([])) total_steps = round(steps * denoise) if steps <= 20: @@ -362,8 +367,16 @@ def get_sigmas(self, coeff, steps, denoise): sigmas = sigmas[-(total_steps + 1):] sigmas[-1] = 0 - return (torch.FloatTensor(sigmas), ) + return io.NodeOutput(torch.FloatTensor(sigmas)) -NODE_CLASS_MAPPINGS = { - "GITSScheduler": GITSScheduler, -} + +class GITSSchedulerExtension(ComfyExtension): + @override + async def get_node_list(self) -> list[type[io.ComfyNode]]: + return [ + GITSScheduler, + ] + + +async def comfy_entrypoint() -> GITSSchedulerExtension: + return GITSSchedulerExtension() From e0210ce0a7140e0c61bce7fdb964b5e5e8d31619 Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Wed, 1 Oct 2025 22:17:33 +0300 Subject: [PATCH 04/37] convert nodes_differential_diffusion.py to V3 schema (#10056) --- comfy_extras/nodes_differential_diffusion.py | 71 +++++++++++--------- 1 file changed, 41 insertions(+), 30 deletions(-) diff --git a/comfy_extras/nodes_differential_diffusion.py b/comfy_extras/nodes_differential_diffusion.py index 255ac420d4eb..6dfdf466c381 100644 --- a/comfy_extras/nodes_differential_diffusion.py +++ b/comfy_extras/nodes_differential_diffusion.py @@ -1,34 +1,41 @@ # code adapted from https://github.com/exx8/differential-diffusion +from typing_extensions import override + import torch +from comfy_api.latest import ComfyExtension, io + + +class DifferentialDiffusion(io.ComfyNode): + @classmethod + def define_schema(cls): + return io.Schema( + node_id="DifferentialDiffusion", + display_name="Differential Diffusion", + category="_for_testing", + inputs=[ + io.Model.Input("model"), + io.Float.Input( + "strength", + default=1.0, + min=0.0, + max=1.0, + step=0.01, + optional=True, + ), + ], + outputs=[io.Model.Output()], + is_experimental=True, + ) -class DifferentialDiffusion(): @classmethod - def INPUT_TYPES(s): - return { - "required": { - "model": ("MODEL", ), - }, - "optional": { - "strength": ("FLOAT", { - "default": 1.0, - "min": 0.0, - "max": 1.0, - "step": 0.01, - }), - } - } - RETURN_TYPES = ("MODEL",) - FUNCTION = "apply" - CATEGORY = "_for_testing" - INIT = False - - def apply(self, model, strength=1.0): + def execute(cls, model, strength=1.0) -> io.NodeOutput: model = model.clone() - model.set_model_denoise_mask_function(lambda *args, **kwargs: self.forward(*args, **kwargs, strength=strength)) - return (model, ) + model.set_model_denoise_mask_function(lambda *args, **kwargs: cls.forward(*args, **kwargs, strength=strength)) + return io.NodeOutput(model) - def forward(self, sigma: torch.Tensor, denoise_mask: torch.Tensor, extra_options: dict, strength: float): + @classmethod + def forward(cls, sigma: torch.Tensor, denoise_mask: torch.Tensor, extra_options: dict, strength: float): model = extra_options["model"] step_sigmas = extra_options["sigmas"] sigma_to = model.inner_model.model_sampling.sigma_min @@ -53,9 +60,13 @@ def forward(self, sigma: torch.Tensor, denoise_mask: torch.Tensor, extra_options return binary_mask -NODE_CLASS_MAPPINGS = { - "DifferentialDiffusion": DifferentialDiffusion, -} -NODE_DISPLAY_NAME_MAPPINGS = { - "DifferentialDiffusion": "Differential Diffusion", -} +class DifferentialDiffusionExtension(ComfyExtension): + @override + async def get_node_list(self) -> list[type[io.ComfyNode]]: + return [ + DifferentialDiffusion, + ] + + +async def comfy_entrypoint() -> DifferentialDiffusionExtension: + return DifferentialDiffusionExtension() From 3af1881455fb0c44c3030b2d61b79302933386d2 Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Wed, 1 Oct 2025 22:18:04 +0300 Subject: [PATCH 05/37] convert nodes_optimalsteps.py to V3 schema (#10074) --- comfy_extras/nodes_optimalsteps.py | 54 +++++++++++++++++++----------- 1 file changed, 34 insertions(+), 20 deletions(-) diff --git a/comfy_extras/nodes_optimalsteps.py b/comfy_extras/nodes_optimalsteps.py index e7c851ca211c..73f0104d8bc5 100644 --- a/comfy_extras/nodes_optimalsteps.py +++ b/comfy_extras/nodes_optimalsteps.py @@ -1,9 +1,12 @@ # from https://github.com/bebebe666/OptimalSteps - import numpy as np import torch +from typing_extensions import override +from comfy_api.latest import ComfyExtension, io + + def loglinear_interp(t_steps, num_steps): """ Performs log-linear interpolation of a given array of decreasing numbers. @@ -23,25 +26,28 @@ def loglinear_interp(t_steps, num_steps): "Chroma": [0.992, 0.99, 0.988, 0.985, 0.982, 0.978, 0.973, 0.968, 0.961, 0.953, 0.943, 0.931, 0.917, 0.9, 0.881, 0.858, 0.832, 0.802, 0.769, 0.731, 0.69, 0.646, 0.599, 0.55, 0.501, 0.451, 0.402, 0.355, 0.311, 0.27, 0.232, 0.199, 0.169, 0.143, 0.12, 0.101, 0.084, 0.07, 0.058, 0.048, 0.001], } -class OptimalStepsScheduler: +class OptimalStepsScheduler(io.ComfyNode): + @classmethod + def define_schema(cls): + return io.Schema( + node_id="OptimalStepsScheduler", + category="sampling/custom_sampling/schedulers", + inputs=[ + io.Combo.Input("model_type", options=["FLUX", "Wan", "Chroma"]), + io.Int.Input("steps", default=20, min=3, max=1000), + io.Float.Input("denoise", default=1.0, min=0.0, max=1.0, step=0.01), + ], + outputs=[ + io.Sigmas.Output(), + ], + ) + @classmethod - def INPUT_TYPES(s): - return {"required": - {"model_type": (["FLUX", "Wan", "Chroma"], ), - "steps": ("INT", {"default": 20, "min": 3, "max": 1000}), - "denoise": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 1.0, "step": 0.01}), - } - } - RETURN_TYPES = ("SIGMAS",) - CATEGORY = "sampling/custom_sampling/schedulers" - - FUNCTION = "get_sigmas" - - def get_sigmas(self, model_type, steps, denoise): + def execute(cls, model_type, steps, denoise) ->io.NodeOutput: total_steps = steps if denoise < 1.0: if denoise <= 0.0: - return (torch.FloatTensor([]),) + return io.NodeOutput(torch.FloatTensor([])) total_steps = round(steps * denoise) sigmas = NOISE_LEVELS[model_type][:] @@ -50,8 +56,16 @@ def get_sigmas(self, model_type, steps, denoise): sigmas = sigmas[-(total_steps + 1):] sigmas[-1] = 0 - return (torch.FloatTensor(sigmas), ) + return io.NodeOutput(torch.FloatTensor(sigmas)) -NODE_CLASS_MAPPINGS = { - "OptimalStepsScheduler": OptimalStepsScheduler, -} + +class OptimalStepsExtension(ComfyExtension): + @override + async def get_node_list(self) -> list[type[io.ComfyNode]]: + return [ + OptimalStepsScheduler, + ] + + +async def comfy_entrypoint() -> OptimalStepsExtension: + return OptimalStepsExtension() From 11bab7be76d0bfdb326e8aea53cdfebd99b42cc5 Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Wed, 1 Oct 2025 22:18:49 +0300 Subject: [PATCH 06/37] convert nodes_pag.py to V3 schema (#10080) --- comfy_extras/nodes_pag.py | 49 +++++++++++++++++++++++++-------------- 1 file changed, 31 insertions(+), 18 deletions(-) diff --git a/comfy_extras/nodes_pag.py b/comfy_extras/nodes_pag.py index eb28196f41c5..79fea5f0cee4 100644 --- a/comfy_extras/nodes_pag.py +++ b/comfy_extras/nodes_pag.py @@ -3,25 +3,30 @@ #My modified one here is more basic but has less chances of breaking with ComfyUI updates. +from typing_extensions import override + import comfy.model_patcher import comfy.samplers +from comfy_api.latest import ComfyExtension, io -class PerturbedAttentionGuidance: - @classmethod - def INPUT_TYPES(s): - return { - "required": { - "model": ("MODEL",), - "scale": ("FLOAT", {"default": 3.0, "min": 0.0, "max": 100.0, "step": 0.01, "round": 0.01}), - } - } - - RETURN_TYPES = ("MODEL",) - FUNCTION = "patch" - CATEGORY = "model_patches/unet" +class PerturbedAttentionGuidance(io.ComfyNode): + @classmethod + def define_schema(cls): + return io.Schema( + node_id="PerturbedAttentionGuidance", + category="model_patches/unet", + inputs=[ + io.Model.Input("model"), + io.Float.Input("scale", default=3.0, min=0.0, max=100.0, step=0.01, round=0.01), + ], + outputs=[ + io.Model.Output(), + ], + ) - def patch(self, model, scale): + @classmethod + def execute(cls, model, scale) -> io.NodeOutput: unet_block = "middle" unet_block_id = 0 m = model.clone() @@ -49,8 +54,16 @@ def post_cfg_function(args): m.set_model_sampler_post_cfg_function(post_cfg_function) - return (m,) + return io.NodeOutput(m) + + +class PAGExtension(ComfyExtension): + @override + async def get_node_list(self) -> list[type[io.ComfyNode]]: + return [ + PerturbedAttentionGuidance, + ] + -NODE_CLASS_MAPPINGS = { - "PerturbedAttentionGuidance": PerturbedAttentionGuidance, -} +async def comfy_entrypoint() -> PAGExtension: + return PAGExtension() From d9c0a4053d955c7fd3400be07001bc4e774591e1 Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Wed, 1 Oct 2025 22:19:56 +0300 Subject: [PATCH 07/37] convert nodes_lt.py to V3 schema (#10084) --- comfy_extras/nodes_lt.py | 424 +++++++++++++++++++++------------------ 1 file changed, 234 insertions(+), 190 deletions(-) diff --git a/comfy_extras/nodes_lt.py b/comfy_extras/nodes_lt.py index f82337a671ed..b51d1580453a 100644 --- a/comfy_extras/nodes_lt.py +++ b/comfy_extras/nodes_lt.py @@ -1,4 +1,3 @@ -import io import nodes import node_helpers import torch @@ -8,46 +7,60 @@ import math import numpy as np import av +from io import BytesIO +from typing_extensions import override from comfy.ldm.lightricks.symmetric_patchifier import SymmetricPatchifier, latent_to_pixel_coords +from comfy_api.latest import ComfyExtension, io -class EmptyLTXVLatentVideo: +class EmptyLTXVLatentVideo(io.ComfyNode): @classmethod - def INPUT_TYPES(s): - return {"required": { "width": ("INT", {"default": 768, "min": 64, "max": nodes.MAX_RESOLUTION, "step": 32}), - "height": ("INT", {"default": 512, "min": 64, "max": nodes.MAX_RESOLUTION, "step": 32}), - "length": ("INT", {"default": 97, "min": 1, "max": nodes.MAX_RESOLUTION, "step": 8}), - "batch_size": ("INT", {"default": 1, "min": 1, "max": 4096})}} - RETURN_TYPES = ("LATENT",) - FUNCTION = "generate" - - CATEGORY = "latent/video/ltxv" + def define_schema(cls): + return io.Schema( + node_id="EmptyLTXVLatentVideo", + category="latent/video/ltxv", + inputs=[ + io.Int.Input("width", default=768, min=64, max=nodes.MAX_RESOLUTION, step=32), + io.Int.Input("height", default=512, min=64, max=nodes.MAX_RESOLUTION, step=32), + io.Int.Input("length", default=97, min=1, max=nodes.MAX_RESOLUTION, step=8), + io.Int.Input("batch_size", default=1, min=1, max=4096), + ], + outputs=[ + io.Latent.Output(), + ], + ) - def generate(self, width, height, length, batch_size=1): + @classmethod + def execute(cls, width, height, length, batch_size=1) -> io.NodeOutput: latent = torch.zeros([batch_size, 128, ((length - 1) // 8) + 1, height // 32, width // 32], device=comfy.model_management.intermediate_device()) - return ({"samples": latent}, ) + return io.NodeOutput({"samples": latent}) -class LTXVImgToVideo: +class LTXVImgToVideo(io.ComfyNode): + @classmethod + def define_schema(cls): + return io.Schema( + node_id="LTXVImgToVideo", + category="conditioning/video_models", + inputs=[ + io.Conditioning.Input("positive"), + io.Conditioning.Input("negative"), + io.Vae.Input("vae"), + io.Image.Input("image"), + io.Int.Input("width", default=768, min=64, max=nodes.MAX_RESOLUTION, step=32), + io.Int.Input("height", default=512, min=64, max=nodes.MAX_RESOLUTION, step=32), + io.Int.Input("length", default=97, min=9, max=nodes.MAX_RESOLUTION, step=8), + io.Int.Input("batch_size", default=1, min=1, max=4096), + io.Float.Input("strength", default=1.0, min=0.0, max=1.0), + ], + outputs=[ + io.Conditioning.Output(display_name="positive"), + io.Conditioning.Output(display_name="negative"), + io.Latent.Output(display_name="latent"), + ], + ) + @classmethod - def INPUT_TYPES(s): - return {"required": {"positive": ("CONDITIONING", ), - "negative": ("CONDITIONING", ), - "vae": ("VAE",), - "image": ("IMAGE",), - "width": ("INT", {"default": 768, "min": 64, "max": nodes.MAX_RESOLUTION, "step": 32}), - "height": ("INT", {"default": 512, "min": 64, "max": nodes.MAX_RESOLUTION, "step": 32}), - "length": ("INT", {"default": 97, "min": 9, "max": nodes.MAX_RESOLUTION, "step": 8}), - "batch_size": ("INT", {"default": 1, "min": 1, "max": 4096}), - "strength": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 1.0}), - }} - - RETURN_TYPES = ("CONDITIONING", "CONDITIONING", "LATENT") - RETURN_NAMES = ("positive", "negative", "latent") - - CATEGORY = "conditioning/video_models" - FUNCTION = "generate" - - def generate(self, positive, negative, image, vae, width, height, length, batch_size, strength): + def execute(cls, positive, negative, image, vae, width, height, length, batch_size, strength) -> io.NodeOutput: pixels = comfy.utils.common_upscale(image.movedim(-1, 1), width, height, "bilinear", "center").movedim(1, -1) encode_pixels = pixels[:, :, :, :3] t = vae.encode(encode_pixels) @@ -62,7 +75,7 @@ def generate(self, positive, negative, image, vae, width, height, length, batch_ ) conditioning_latent_frames_mask[:, :, :t.shape[2]] = 1.0 - strength - return (positive, negative, {"samples": latent, "noise_mask": conditioning_latent_frames_mask}, ) + return io.NodeOutput(positive, negative, {"samples": latent, "noise_mask": conditioning_latent_frames_mask}) def conditioning_get_any_value(conditioning, key, default=None): @@ -93,35 +106,46 @@ def get_keyframe_idxs(cond): num_keyframes = torch.unique(keyframe_idxs[:, 0]).shape[0] return keyframe_idxs, num_keyframes -class LTXVAddGuide: +class LTXVAddGuide(io.ComfyNode): + NUM_PREFIX_FRAMES = 2 + PATCHIFIER = SymmetricPatchifier(1) + @classmethod - def INPUT_TYPES(s): - return {"required": {"positive": ("CONDITIONING", ), - "negative": ("CONDITIONING", ), - "vae": ("VAE",), - "latent": ("LATENT",), - "image": ("IMAGE", {"tooltip": "Image or video to condition the latent video on. Must be 8*n + 1 frames." - "If the video is not 8*n + 1 frames, it will be cropped to the nearest 8*n + 1 frames."}), - "frame_idx": ("INT", {"default": 0, "min": -9999, "max": 9999, - "tooltip": "Frame index to start the conditioning at. For single-frame images or " - "videos with 1-8 frames, any frame_idx value is acceptable. For videos with 9+ " - "frames, frame_idx must be divisible by 8, otherwise it will be rounded down to " - "the nearest multiple of 8. Negative values are counted from the end of the video."}), - "strength": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 1.0, "step": 0.01}), - } - } - - RETURN_TYPES = ("CONDITIONING", "CONDITIONING", "LATENT") - RETURN_NAMES = ("positive", "negative", "latent") - - CATEGORY = "conditioning/video_models" - FUNCTION = "generate" - - def __init__(self): - self._num_prefix_frames = 2 - self._patchifier = SymmetricPatchifier(1) - - def encode(self, vae, latent_width, latent_height, images, scale_factors): + def define_schema(cls): + return io.Schema( + node_id="LTXVAddGuide", + category="conditioning/video_models", + inputs=[ + io.Conditioning.Input("positive"), + io.Conditioning.Input("negative"), + io.Vae.Input("vae"), + io.Latent.Input("latent"), + io.Image.Input( + "image", + tooltip="Image or video to condition the latent video on. Must be 8*n + 1 frames. " + "If the video is not 8*n + 1 frames, it will be cropped to the nearest 8*n + 1 frames.", + ), + io.Int.Input( + "frame_idx", + default=0, + min=-9999, + max=9999, + tooltip="Frame index to start the conditioning at. " + "For single-frame images or videos with 1-8 frames, any frame_idx value is acceptable. " + "For videos with 9+ frames, frame_idx must be divisible by 8, otherwise it will be rounded " + "down to the nearest multiple of 8. Negative values are counted from the end of the video.", + ), + io.Float.Input("strength", default=1.0, min=0.0, max=1.0, step=0.01), + ], + outputs=[ + io.Conditioning.Output(display_name="positive"), + io.Conditioning.Output(display_name="negative"), + io.Latent.Output(display_name="latent"), + ], + ) + + @classmethod + def encode(cls, vae, latent_width, latent_height, images, scale_factors): time_scale_factor, width_scale_factor, height_scale_factor = scale_factors images = images[:(images.shape[0] - 1) // time_scale_factor * time_scale_factor + 1] pixels = comfy.utils.common_upscale(images.movedim(-1, 1), latent_width * width_scale_factor, latent_height * height_scale_factor, "bilinear", crop="disabled").movedim(1, -1) @@ -129,7 +153,8 @@ def encode(self, vae, latent_width, latent_height, images, scale_factors): t = vae.encode(encode_pixels) return encode_pixels, t - def get_latent_index(self, cond, latent_length, guide_length, frame_idx, scale_factors): + @classmethod + def get_latent_index(cls, cond, latent_length, guide_length, frame_idx, scale_factors): time_scale_factor, _, _ = scale_factors _, num_keyframes = get_keyframe_idxs(cond) latent_count = latent_length - num_keyframes @@ -141,9 +166,10 @@ def get_latent_index(self, cond, latent_length, guide_length, frame_idx, scale_f return frame_idx, latent_idx - def add_keyframe_index(self, cond, frame_idx, guiding_latent, scale_factors): + @classmethod + def add_keyframe_index(cls, cond, frame_idx, guiding_latent, scale_factors): keyframe_idxs, _ = get_keyframe_idxs(cond) - _, latent_coords = self._patchifier.patchify(guiding_latent) + _, latent_coords = cls.PATCHIFIER.patchify(guiding_latent) pixel_coords = latent_to_pixel_coords(latent_coords, scale_factors, causal_fix=frame_idx == 0) # we need the causal fix only if we're placing the new latents at index 0 pixel_coords[:, 0] += frame_idx if keyframe_idxs is None: @@ -152,8 +178,9 @@ def add_keyframe_index(self, cond, frame_idx, guiding_latent, scale_factors): keyframe_idxs = torch.cat([keyframe_idxs, pixel_coords], dim=2) return node_helpers.conditioning_set_values(cond, {"keyframe_idxs": keyframe_idxs}) - def append_keyframe(self, positive, negative, frame_idx, latent_image, noise_mask, guiding_latent, strength, scale_factors): - _, latent_idx = self.get_latent_index( + @classmethod + def append_keyframe(cls, positive, negative, frame_idx, latent_image, noise_mask, guiding_latent, strength, scale_factors): + _, latent_idx = cls.get_latent_index( cond=positive, latent_length=latent_image.shape[2], guide_length=guiding_latent.shape[2], @@ -162,8 +189,8 @@ def append_keyframe(self, positive, negative, frame_idx, latent_image, noise_mas ) noise_mask[:, :, latent_idx:latent_idx + guiding_latent.shape[2]] = 1.0 - positive = self.add_keyframe_index(positive, frame_idx, guiding_latent, scale_factors) - negative = self.add_keyframe_index(negative, frame_idx, guiding_latent, scale_factors) + positive = cls.add_keyframe_index(positive, frame_idx, guiding_latent, scale_factors) + negative = cls.add_keyframe_index(negative, frame_idx, guiding_latent, scale_factors) mask = torch.full( (noise_mask.shape[0], 1, guiding_latent.shape[2], noise_mask.shape[3], noise_mask.shape[4]), @@ -176,7 +203,8 @@ def append_keyframe(self, positive, negative, frame_idx, latent_image, noise_mas noise_mask = torch.cat([noise_mask, mask], dim=2) return positive, negative, latent_image, noise_mask - def replace_latent_frames(self, latent_image, noise_mask, guiding_latent, latent_idx, strength): + @classmethod + def replace_latent_frames(cls, latent_image, noise_mask, guiding_latent, latent_idx, strength): cond_length = guiding_latent.shape[2] assert latent_image.shape[2] >= latent_idx + cond_length, "Conditioning frames exceed the length of the latent sequence." @@ -195,20 +223,21 @@ def replace_latent_frames(self, latent_image, noise_mask, guiding_latent, latent return latent_image, noise_mask - def generate(self, positive, negative, vae, latent, image, frame_idx, strength): + @classmethod + def execute(cls, positive, negative, vae, latent, image, frame_idx, strength) -> io.NodeOutput: scale_factors = vae.downscale_index_formula latent_image = latent["samples"] noise_mask = get_noise_mask(latent) _, _, latent_length, latent_height, latent_width = latent_image.shape - image, t = self.encode(vae, latent_width, latent_height, image, scale_factors) + image, t = cls.encode(vae, latent_width, latent_height, image, scale_factors) - frame_idx, latent_idx = self.get_latent_index(positive, latent_length, len(image), frame_idx, scale_factors) + frame_idx, latent_idx = cls.get_latent_index(positive, latent_length, len(image), frame_idx, scale_factors) assert latent_idx + t.shape[2] <= latent_length, "Conditioning frames exceed the length of the latent sequence." - num_prefix_frames = min(self._num_prefix_frames, t.shape[2]) + num_prefix_frames = min(cls.NUM_PREFIX_FRAMES, t.shape[2]) - positive, negative, latent_image, noise_mask = self.append_keyframe( + positive, negative, latent_image, noise_mask = cls.append_keyframe( positive, negative, frame_idx, @@ -223,9 +252,9 @@ def generate(self, positive, negative, vae, latent, image, frame_idx, strength): t = t[:, :, num_prefix_frames:] if t.shape[2] == 0: - return (positive, negative, {"samples": latent_image, "noise_mask": noise_mask},) + return io.NodeOutput(positive, negative, {"samples": latent_image, "noise_mask": noise_mask}) - latent_image, noise_mask = self.replace_latent_frames( + latent_image, noise_mask = cls.replace_latent_frames( latent_image, noise_mask, t, @@ -233,34 +262,35 @@ def generate(self, positive, negative, vae, latent, image, frame_idx, strength): strength, ) - return (positive, negative, {"samples": latent_image, "noise_mask": noise_mask},) + return io.NodeOutput(positive, negative, {"samples": latent_image, "noise_mask": noise_mask}) -class LTXVCropGuides: +class LTXVCropGuides(io.ComfyNode): @classmethod - def INPUT_TYPES(s): - return {"required": {"positive": ("CONDITIONING", ), - "negative": ("CONDITIONING", ), - "latent": ("LATENT",), - } - } - - RETURN_TYPES = ("CONDITIONING", "CONDITIONING", "LATENT") - RETURN_NAMES = ("positive", "negative", "latent") - - CATEGORY = "conditioning/video_models" - FUNCTION = "crop" - - def __init__(self): - self._patchifier = SymmetricPatchifier(1) + def define_schema(cls): + return io.Schema( + node_id="LTXVCropGuides", + category="conditioning/video_models", + inputs=[ + io.Conditioning.Input("positive"), + io.Conditioning.Input("negative"), + io.Latent.Input("latent"), + ], + outputs=[ + io.Conditioning.Output(display_name="positive"), + io.Conditioning.Output(display_name="negative"), + io.Latent.Output(display_name="latent"), + ], + ) - def crop(self, positive, negative, latent): + @classmethod + def execute(cls, positive, negative, latent) -> io.NodeOutput: latent_image = latent["samples"].clone() noise_mask = get_noise_mask(latent) _, num_keyframes = get_keyframe_idxs(positive) if num_keyframes == 0: - return (positive, negative, {"samples": latent_image, "noise_mask": noise_mask},) + return io.NodeOutput(positive, negative, {"samples": latent_image, "noise_mask": noise_mask},) latent_image = latent_image[:, :, :-num_keyframes] noise_mask = noise_mask[:, :, :-num_keyframes] @@ -268,44 +298,52 @@ def crop(self, positive, negative, latent): positive = node_helpers.conditioning_set_values(positive, {"keyframe_idxs": None}) negative = node_helpers.conditioning_set_values(negative, {"keyframe_idxs": None}) - return (positive, negative, {"samples": latent_image, "noise_mask": noise_mask},) + return io.NodeOutput(positive, negative, {"samples": latent_image, "noise_mask": noise_mask}) -class LTXVConditioning: +class LTXVConditioning(io.ComfyNode): @classmethod - def INPUT_TYPES(s): - return {"required": {"positive": ("CONDITIONING", ), - "negative": ("CONDITIONING", ), - "frame_rate": ("FLOAT", {"default": 25.0, "min": 0.0, "max": 1000.0, "step": 0.01}), - }} - RETURN_TYPES = ("CONDITIONING", "CONDITIONING") - RETURN_NAMES = ("positive", "negative") - FUNCTION = "append" - - CATEGORY = "conditioning/video_models" - - def append(self, positive, negative, frame_rate): + def define_schema(cls): + return io.Schema( + node_id="LTXVConditioning", + category="conditioning/video_models", + inputs=[ + io.Conditioning.Input("positive"), + io.Conditioning.Input("negative"), + io.Float.Input("frame_rate", default=25.0, min=0.0, max=1000.0, step=0.01), + ], + outputs=[ + io.Conditioning.Output(display_name="positive"), + io.Conditioning.Output(display_name="negative"), + ], + ) + + @classmethod + def execute(cls, positive, negative, frame_rate) -> io.NodeOutput: positive = node_helpers.conditioning_set_values(positive, {"frame_rate": frame_rate}) negative = node_helpers.conditioning_set_values(negative, {"frame_rate": frame_rate}) - return (positive, negative) + return io.NodeOutput(positive, negative) -class ModelSamplingLTXV: +class ModelSamplingLTXV(io.ComfyNode): @classmethod - def INPUT_TYPES(s): - return {"required": { "model": ("MODEL",), - "max_shift": ("FLOAT", {"default": 2.05, "min": 0.0, "max": 100.0, "step":0.01}), - "base_shift": ("FLOAT", {"default": 0.95, "min": 0.0, "max": 100.0, "step":0.01}), - }, - "optional": {"latent": ("LATENT",), } - } - - RETURN_TYPES = ("MODEL",) - FUNCTION = "patch" - - CATEGORY = "advanced/model" + def define_schema(cls): + return io.Schema( + node_id="ModelSamplingLTXV", + category="advanced/model", + inputs=[ + io.Model.Input("model"), + io.Float.Input("max_shift", default=2.05, min=0.0, max=100.0, step=0.01), + io.Float.Input("base_shift", default=0.95, min=0.0, max=100.0, step=0.01), + io.Latent.Input("latent", optional=True), + ], + outputs=[ + io.Model.Output(), + ], + ) - def patch(self, model, max_shift, base_shift, latent=None): + @classmethod + def execute(cls, model, max_shift, base_shift, latent=None) -> io.NodeOutput: m = model.clone() if latent is None: @@ -329,37 +367,41 @@ class ModelSamplingAdvanced(sampling_base, sampling_type): model_sampling.set_parameters(shift=shift) m.add_object_patch("model_sampling", model_sampling) - return (m, ) + return io.NodeOutput(m) -class LTXVScheduler: +class LTXVScheduler(io.ComfyNode): @classmethod - def INPUT_TYPES(s): - return {"required": - {"steps": ("INT", {"default": 20, "min": 1, "max": 10000}), - "max_shift": ("FLOAT", {"default": 2.05, "min": 0.0, "max": 100.0, "step":0.01}), - "base_shift": ("FLOAT", {"default": 0.95, "min": 0.0, "max": 100.0, "step":0.01}), - "stretch": ("BOOLEAN", { - "default": True, - "tooltip": "Stretch the sigmas to be in the range [terminal, 1]." - }), - "terminal": ( - "FLOAT", - { - "default": 0.1, "min": 0.0, "max": 0.99, "step": 0.01, - "tooltip": "The terminal value of the sigmas after stretching." - }, - ), - }, - "optional": {"latent": ("LATENT",), } - } - - RETURN_TYPES = ("SIGMAS",) - CATEGORY = "sampling/custom_sampling/schedulers" - - FUNCTION = "get_sigmas" - - def get_sigmas(self, steps, max_shift, base_shift, stretch, terminal, latent=None): + def define_schema(cls): + return io.Schema( + node_id="LTXVScheduler", + category="sampling/custom_sampling/schedulers", + inputs=[ + io.Int.Input("steps", default=20, min=1, max=10000), + io.Float.Input("max_shift", default=2.05, min=0.0, max=100.0, step=0.01), + io.Float.Input("base_shift", default=0.95, min=0.0, max=100.0, step=0.01), + io.Boolean.Input( + id="stretch", + default=True, + tooltip="Stretch the sigmas to be in the range [terminal, 1].", + ), + io.Float.Input( + id="terminal", + default=0.1, + min=0.0, + max=0.99, + step=0.01, + tooltip="The terminal value of the sigmas after stretching.", + ), + io.Latent.Input("latent", optional=True), + ], + outputs=[ + io.Sigmas.Output(), + ], + ) + + @classmethod + def execute(cls, steps, max_shift, base_shift, stretch, terminal, latent=None) -> io.NodeOutput: if latent is None: tokens = 4096 else: @@ -389,7 +431,7 @@ def get_sigmas(self, steps, max_shift, base_shift, stretch, terminal, latent=Non stretched = 1.0 - (one_minus_z / scale_factor) sigmas[non_zero_mask] = stretched - return (sigmas,) + return io.NodeOutput(sigmas) def encode_single_frame(output_file, image_array: np.ndarray, crf): container = av.open(output_file, "w", format="mp4") @@ -423,52 +465,54 @@ def preprocess(image: torch.Tensor, crf=29): return image image_array = (image[:(image.shape[0] // 2) * 2, :(image.shape[1] // 2) * 2] * 255.0).byte().cpu().numpy() - with io.BytesIO() as output_file: + with BytesIO() as output_file: encode_single_frame(output_file, image_array, crf) video_bytes = output_file.getvalue() - with io.BytesIO(video_bytes) as video_file: + with BytesIO(video_bytes) as video_file: image_array = decode_single_frame(video_file) tensor = torch.tensor(image_array, dtype=image.dtype, device=image.device) / 255.0 return tensor -class LTXVPreprocess: +class LTXVPreprocess(io.ComfyNode): @classmethod - def INPUT_TYPES(s): - return { - "required": { - "image": ("IMAGE",), - "img_compression": ( - "INT", - { - "default": 35, - "min": 0, - "max": 100, - "tooltip": "Amount of compression to apply on image.", - }, + def define_schema(cls): + return io.Schema( + node_id="LTXVPreprocess", + category="image", + inputs=[ + io.Image.Input("image"), + io.Int.Input( + id="img_compression", default=35, min=0, max=100, tooltip="Amount of compression to apply on image." ), - } - } - - FUNCTION = "preprocess" - RETURN_TYPES = ("IMAGE",) - RETURN_NAMES = ("output_image",) - CATEGORY = "image" + ], + outputs=[ + io.Image.Output(display_name="output_image"), + ], + ) - def preprocess(self, image, img_compression): + @classmethod + def execute(cls, image, img_compression) -> io.NodeOutput: output_images = [] for i in range(image.shape[0]): output_images.append(preprocess(image[i], img_compression)) - return (torch.stack(output_images),) - - -NODE_CLASS_MAPPINGS = { - "EmptyLTXVLatentVideo": EmptyLTXVLatentVideo, - "LTXVImgToVideo": LTXVImgToVideo, - "ModelSamplingLTXV": ModelSamplingLTXV, - "LTXVConditioning": LTXVConditioning, - "LTXVScheduler": LTXVScheduler, - "LTXVAddGuide": LTXVAddGuide, - "LTXVPreprocess": LTXVPreprocess, - "LTXVCropGuides": LTXVCropGuides, -} + return io.NodeOutput(torch.stack(output_images)) + + +class LtxvExtension(ComfyExtension): + @override + async def get_node_list(self) -> list[type[io.ComfyNode]]: + return [ + EmptyLTXVLatentVideo, + LTXVImgToVideo, + ModelSamplingLTXV, + LTXVConditioning, + LTXVScheduler, + LTXVAddGuide, + LTXVPreprocess, + LTXVCropGuides, + ] + + +async def comfy_entrypoint() -> LtxvExtension: + return LtxvExtension() From e4f99b479a19730bea890567129f4032b4dd4787 Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Wed, 1 Oct 2025 22:20:30 +0300 Subject: [PATCH 08/37] convert nodes_ip2p.pt to V3 schema (#10097) --- comfy_extras/nodes_ip2p.py | 52 +++++++++++++++++++++++++------------- 1 file changed, 35 insertions(+), 17 deletions(-) diff --git a/comfy_extras/nodes_ip2p.py b/comfy_extras/nodes_ip2p.py index c2e70a84c10c..78f29915db2e 100644 --- a/comfy_extras/nodes_ip2p.py +++ b/comfy_extras/nodes_ip2p.py @@ -1,21 +1,30 @@ import torch -class InstructPixToPixConditioning: - @classmethod - def INPUT_TYPES(s): - return {"required": {"positive": ("CONDITIONING", ), - "negative": ("CONDITIONING", ), - "vae": ("VAE", ), - "pixels": ("IMAGE", ), - }} +from typing_extensions import override +from comfy_api.latest import ComfyExtension, io - RETURN_TYPES = ("CONDITIONING","CONDITIONING","LATENT") - RETURN_NAMES = ("positive", "negative", "latent") - FUNCTION = "encode" - CATEGORY = "conditioning/instructpix2pix" +class InstructPixToPixConditioning(io.ComfyNode): + @classmethod + def define_schema(cls): + return io.Schema( + node_id="InstructPixToPixConditioning", + category="conditioning/instructpix2pix", + inputs=[ + io.Conditioning.Input("positive"), + io.Conditioning.Input("negative"), + io.Vae.Input("vae"), + io.Image.Input("pixels"), + ], + outputs=[ + io.Conditioning.Output(display_name="positive"), + io.Conditioning.Output(display_name="negative"), + io.Latent.Output(display_name="latent"), + ], + ) - def encode(self, positive, negative, pixels, vae): + @classmethod + def execute(cls, positive, negative, pixels, vae) -> io.NodeOutput: x = (pixels.shape[1] // 8) * 8 y = (pixels.shape[2] // 8) * 8 @@ -38,8 +47,17 @@ def encode(self, positive, negative, pixels, vae): n = [t[0], d] c.append(n) out.append(c) - return (out[0], out[1], out_latent) + return io.NodeOutput(out[0], out[1], out_latent) + + +class InstructPix2PixExtension(ComfyExtension): + @override + async def get_node_list(self) -> list[type[io.ComfyNode]]: + return [ + InstructPixToPixConditioning, + ] + + +async def comfy_entrypoint() -> InstructPix2PixExtension: + return InstructPix2PixExtension() -NODE_CLASS_MAPPINGS = { - "InstructPixToPixConditioning": InstructPixToPixConditioning, -} From a6f83a4a1a70d720c16d66feb5d87fee5998acdf Mon Sep 17 00:00:00 2001 From: comfyanonymous <121283862+comfyanonymous@users.noreply.github.com> Date: Wed, 1 Oct 2025 14:19:13 -0700 Subject: [PATCH 09/37] Support the new hunyuan vae. (#10150) --- comfy/ldm/hunyuan_video/vae_refiner.py | 112 ++++++++++++++++--------- comfy/sd.py | 68 +++++++++------ 2 files changed, 115 insertions(+), 65 deletions(-) diff --git a/comfy/ldm/hunyuan_video/vae_refiner.py b/comfy/ldm/hunyuan_video/vae_refiner.py index c6f742710ca4..c2a0b507d4bf 100644 --- a/comfy/ldm/hunyuan_video/vae_refiner.py +++ b/comfy/ldm/hunyuan_video/vae_refiner.py @@ -1,7 +1,7 @@ import torch import torch.nn as nn import torch.nn.functional as F -from comfy.ldm.modules.diffusionmodules.model import ResnetBlock, AttnBlock, VideoConv3d +from comfy.ldm.modules.diffusionmodules.model import ResnetBlock, AttnBlock, VideoConv3d, Normalize import comfy.ops import comfy.ldm.models.autoencoder ops = comfy.ops.disable_weight_init @@ -17,11 +17,12 @@ def forward(self, x): return F.normalize(x, dim=1) * self.scale * self.gamma class DnSmpl(nn.Module): - def __init__(self, ic, oc, tds=True): + def __init__(self, ic, oc, tds=True, refiner_vae=True, op=VideoConv3d): super().__init__() fct = 2 * 2 * 2 if tds else 1 * 2 * 2 assert oc % fct == 0 - self.conv = VideoConv3d(ic, oc // fct, kernel_size=3) + self.conv = op(ic, oc // fct, kernel_size=3, stride=1, padding=1) + self.refiner_vae = refiner_vae self.tds = tds self.gs = fct * ic // oc @@ -30,7 +31,7 @@ def forward(self, x): r1 = 2 if self.tds else 1 h = self.conv(x) - if self.tds: + if self.tds and self.refiner_vae: hf = h[:, :, :1, :, :] b, c, f, ht, wd = hf.shape hf = hf.reshape(b, c, f, ht // 2, 2, wd // 2, 2) @@ -66,6 +67,7 @@ def forward(self, x): sc = torch.cat([xf, xn], dim=2) else: b, c, frms, ht, wd = h.shape + nf = frms // r1 h = h.reshape(b, c, nf, r1, ht // 2, 2, wd // 2, 2) h = h.permute(0, 3, 5, 7, 1, 2, 4, 6) @@ -83,10 +85,11 @@ def forward(self, x): class UpSmpl(nn.Module): - def __init__(self, ic, oc, tus=True): + def __init__(self, ic, oc, tus=True, refiner_vae=True, op=VideoConv3d): super().__init__() fct = 2 * 2 * 2 if tus else 1 * 2 * 2 - self.conv = VideoConv3d(ic, oc * fct, kernel_size=3) + self.conv = op(ic, oc * fct, kernel_size=3, stride=1, padding=1) + self.refiner_vae = refiner_vae self.tus = tus self.rp = fct * oc // ic @@ -95,7 +98,7 @@ def forward(self, x): r1 = 2 if self.tus else 1 h = self.conv(x) - if self.tus: + if self.tus and self.refiner_vae: hf = h[:, :, :1, :, :] b, c, f, ht, wd = hf.shape nc = c // (2 * 2) @@ -148,43 +151,56 @@ def forward(self, x): class Encoder(nn.Module): def __init__(self, in_channels, z_channels, block_out_channels, num_res_blocks, - ffactor_spatial, ffactor_temporal, downsample_match_channel=True, **_): + ffactor_spatial, ffactor_temporal, downsample_match_channel=True, refiner_vae=True, **_): super().__init__() self.z_channels = z_channels self.block_out_channels = block_out_channels self.num_res_blocks = num_res_blocks - self.conv_in = VideoConv3d(in_channels, block_out_channels[0], 3, 1, 1) + self.ffactor_temporal = ffactor_temporal + + self.refiner_vae = refiner_vae + if self.refiner_vae: + conv_op = VideoConv3d + norm_op = RMS_norm + else: + conv_op = ops.Conv3d + norm_op = Normalize + + self.conv_in = conv_op(in_channels, block_out_channels[0], 3, 1, 1) self.down = nn.ModuleList() ch = block_out_channels[0] depth = (ffactor_spatial >> 1).bit_length() - depth_temporal = ((ffactor_spatial // ffactor_temporal) >> 1).bit_length() + depth_temporal = ((ffactor_spatial // self.ffactor_temporal) >> 1).bit_length() for i, tgt in enumerate(block_out_channels): stage = nn.Module() stage.block = nn.ModuleList([ResnetBlock(in_channels=ch if j == 0 else tgt, out_channels=tgt, temb_channels=0, - conv_op=VideoConv3d, norm_op=RMS_norm) + conv_op=conv_op, norm_op=norm_op) for j in range(num_res_blocks)]) ch = tgt if i < depth: nxt = block_out_channels[i + 1] if i + 1 < len(block_out_channels) and downsample_match_channel else ch - stage.downsample = DnSmpl(ch, nxt, tds=i >= depth_temporal) + stage.downsample = DnSmpl(ch, nxt, tds=i >= depth_temporal, refiner_vae=self.refiner_vae, op=conv_op) ch = nxt self.down.append(stage) self.mid = nn.Module() - self.mid.block_1 = ResnetBlock(in_channels=ch, out_channels=ch, temb_channels=0, conv_op=VideoConv3d, norm_op=RMS_norm) - self.mid.attn_1 = AttnBlock(ch, conv_op=ops.Conv3d, norm_op=RMS_norm) - self.mid.block_2 = ResnetBlock(in_channels=ch, out_channels=ch, temb_channels=0, conv_op=VideoConv3d, norm_op=RMS_norm) + self.mid.block_1 = ResnetBlock(in_channels=ch, out_channels=ch, temb_channels=0, conv_op=conv_op, norm_op=norm_op) + self.mid.attn_1 = AttnBlock(ch, conv_op=ops.Conv3d, norm_op=norm_op) + self.mid.block_2 = ResnetBlock(in_channels=ch, out_channels=ch, temb_channels=0, conv_op=conv_op, norm_op=norm_op) - self.norm_out = RMS_norm(ch) - self.conv_out = VideoConv3d(ch, z_channels << 1, 3, 1, 1) + self.norm_out = norm_op(ch) + self.conv_out = conv_op(ch, z_channels << 1, 3, 1, 1) self.regul = comfy.ldm.models.autoencoder.DiagonalGaussianRegularizer() def forward(self, x): + if not self.refiner_vae and x.shape[2] == 1: + x = x.expand(-1, -1, self.ffactor_temporal, -1, -1) + x = self.conv_in(x) for stage in self.down: @@ -200,31 +216,42 @@ def forward(self, x): skip = x.view(b, c // grp, grp, t, h, w).mean(2) out = self.conv_out(F.silu(self.norm_out(x))) + skip - out = self.regul(out)[0] - out = torch.cat((out[:, :, :1], out), dim=2) - out = out.permute(0, 2, 1, 3, 4) - b, f_times_2, c, h, w = out.shape - out = out.reshape(b, f_times_2 // 2, 2 * c, h, w) - out = out.permute(0, 2, 1, 3, 4).contiguous() + if self.refiner_vae: + out = self.regul(out)[0] + + out = torch.cat((out[:, :, :1], out), dim=2) + out = out.permute(0, 2, 1, 3, 4) + b, f_times_2, c, h, w = out.shape + out = out.reshape(b, f_times_2 // 2, 2 * c, h, w) + out = out.permute(0, 2, 1, 3, 4).contiguous() + return out class Decoder(nn.Module): def __init__(self, z_channels, out_channels, block_out_channels, num_res_blocks, - ffactor_spatial, ffactor_temporal, upsample_match_channel=True, **_): + ffactor_spatial, ffactor_temporal, upsample_match_channel=True, refiner_vae=True, **_): super().__init__() block_out_channels = block_out_channels[::-1] self.z_channels = z_channels self.block_out_channels = block_out_channels self.num_res_blocks = num_res_blocks + self.refiner_vae = refiner_vae + if self.refiner_vae: + conv_op = VideoConv3d + norm_op = RMS_norm + else: + conv_op = ops.Conv3d + norm_op = Normalize + ch = block_out_channels[0] - self.conv_in = VideoConv3d(z_channels, ch, 3) + self.conv_in = conv_op(z_channels, ch, kernel_size=3, stride=1, padding=1) self.mid = nn.Module() - self.mid.block_1 = ResnetBlock(in_channels=ch, out_channels=ch, temb_channels=0, conv_op=VideoConv3d, norm_op=RMS_norm) - self.mid.attn_1 = AttnBlock(ch, conv_op=ops.Conv3d, norm_op=RMS_norm) - self.mid.block_2 = ResnetBlock(in_channels=ch, out_channels=ch, temb_channels=0, conv_op=VideoConv3d, norm_op=RMS_norm) + self.mid.block_1 = ResnetBlock(in_channels=ch, out_channels=ch, temb_channels=0, conv_op=conv_op, norm_op=norm_op) + self.mid.attn_1 = AttnBlock(ch, conv_op=ops.Conv3d, norm_op=norm_op) + self.mid.block_2 = ResnetBlock(in_channels=ch, out_channels=ch, temb_channels=0, conv_op=conv_op, norm_op=norm_op) self.up = nn.ModuleList() depth = (ffactor_spatial >> 1).bit_length() @@ -235,25 +262,26 @@ def __init__(self, z_channels, out_channels, block_out_channels, num_res_blocks, stage.block = nn.ModuleList([ResnetBlock(in_channels=ch if j == 0 else tgt, out_channels=tgt, temb_channels=0, - conv_op=VideoConv3d, norm_op=RMS_norm) + conv_op=conv_op, norm_op=norm_op) for j in range(num_res_blocks + 1)]) ch = tgt if i < depth: nxt = block_out_channels[i + 1] if i + 1 < len(block_out_channels) and upsample_match_channel else ch - stage.upsample = UpSmpl(ch, nxt, tus=i < depth_temporal) + stage.upsample = UpSmpl(ch, nxt, tus=i < depth_temporal, refiner_vae=self.refiner_vae, op=conv_op) ch = nxt self.up.append(stage) - self.norm_out = RMS_norm(ch) - self.conv_out = VideoConv3d(ch, out_channels, 3) + self.norm_out = norm_op(ch) + self.conv_out = conv_op(ch, out_channels, 3, stride=1, padding=1) def forward(self, z): - z = z.permute(0, 2, 1, 3, 4) - b, f, c, h, w = z.shape - z = z.reshape(b, f, 2, c // 2, h, w) - z = z.permute(0, 1, 2, 3, 4, 5).reshape(b, f * 2, c // 2, h, w) - z = z.permute(0, 2, 1, 3, 4) - z = z[:, :, 1:] + if self.refiner_vae: + z = z.permute(0, 2, 1, 3, 4) + b, f, c, h, w = z.shape + z = z.reshape(b, f, 2, c // 2, h, w) + z = z.permute(0, 1, 2, 3, 4, 5).reshape(b, f * 2, c // 2, h, w) + z = z.permute(0, 2, 1, 3, 4) + z = z[:, :, 1:] x = self.conv_in(z) + z.repeat_interleave(self.block_out_channels[0] // self.z_channels, 1) x = self.mid.block_2(self.mid.attn_1(self.mid.block_1(x))) @@ -264,4 +292,10 @@ def forward(self, z): if hasattr(stage, 'upsample'): x = stage.upsample(x) - return self.conv_out(F.silu(self.norm_out(x))) + out = self.conv_out(F.silu(self.norm_out(x))) + + if not self.refiner_vae: + if z.shape[-3] == 1: + out = out[:, :, -1:] + + return out diff --git a/comfy/sd.py b/comfy/sd.py index 2df340739f4e..873ad20f2b37 100644 --- a/comfy/sd.py +++ b/comfy/sd.py @@ -332,35 +332,51 @@ def __init__(self, sd=None, device=None, config=None, dtype=None, metadata=None) self.first_stage_model = StageC_coder() self.downscale_ratio = 32 self.latent_channels = 16 - elif "decoder.conv_in.weight" in sd and sd['decoder.conv_in.weight'].shape[1] == 64: - ddconfig = {"block_out_channels": [128, 256, 512, 512, 1024, 1024], "in_channels": 3, "out_channels": 3, "num_res_blocks": 2, "ffactor_spatial": 32, "downsample_match_channel": True, "upsample_match_channel": True} - self.latent_channels = ddconfig['z_channels'] = sd["decoder.conv_in.weight"].shape[1] - self.downscale_ratio = 32 - self.upscale_ratio = 32 - self.working_dtypes = [torch.float16, torch.bfloat16, torch.float32] - self.first_stage_model = AutoencodingEngine(regularizer_config={'target': "comfy.ldm.models.autoencoder.DiagonalGaussianRegularizer"}, - encoder_config={'target': "comfy.ldm.hunyuan_video.vae.Encoder", 'params': ddconfig}, - decoder_config={'target': "comfy.ldm.hunyuan_video.vae.Decoder", 'params': ddconfig}) - - self.memory_used_encode = lambda shape, dtype: (700 * shape[2] * shape[3]) * model_management.dtype_size(dtype) - self.memory_used_decode = lambda shape, dtype: (700 * shape[2] * shape[3] * 32 * 32) * model_management.dtype_size(dtype) - elif "decoder.conv_in.weight" in sd: - #default SD1.x/SD2.x VAE parameters - ddconfig = {'double_z': True, 'z_channels': 4, 'resolution': 256, 'in_channels': 3, 'out_ch': 3, 'ch': 128, 'ch_mult': [1, 2, 4, 4], 'num_res_blocks': 2, 'attn_resolutions': [], 'dropout': 0.0} - - if 'encoder.down.2.downsample.conv.weight' not in sd and 'decoder.up.3.upsample.conv.weight' not in sd: #Stable diffusion x4 upscaler VAE - ddconfig['ch_mult'] = [1, 2, 4] - self.downscale_ratio = 4 - self.upscale_ratio = 4 + if sd['decoder.conv_in.weight'].shape[1] == 64: + ddconfig = {"block_out_channels": [128, 256, 512, 512, 1024, 1024], "in_channels": 3, "out_channels": 3, "num_res_blocks": 2, "ffactor_spatial": 32, "downsample_match_channel": True, "upsample_match_channel": True} + self.latent_channels = ddconfig['z_channels'] = sd["decoder.conv_in.weight"].shape[1] + self.downscale_ratio = 32 + self.upscale_ratio = 32 + self.working_dtypes = [torch.float16, torch.bfloat16, torch.float32] + self.first_stage_model = AutoencodingEngine(regularizer_config={'target': "comfy.ldm.models.autoencoder.DiagonalGaussianRegularizer"}, + encoder_config={'target': "comfy.ldm.hunyuan_video.vae.Encoder", 'params': ddconfig}, + decoder_config={'target': "comfy.ldm.hunyuan_video.vae.Decoder", 'params': ddconfig}) + + self.memory_used_encode = lambda shape, dtype: (700 * shape[2] * shape[3]) * model_management.dtype_size(dtype) + self.memory_used_decode = lambda shape, dtype: (700 * shape[2] * shape[3] * 32 * 32) * model_management.dtype_size(dtype) + elif sd['decoder.conv_in.weight'].shape[1] == 32: + ddconfig = {"block_out_channels": [128, 256, 512, 1024, 1024], "in_channels": 3, "out_channels": 3, "num_res_blocks": 2, "ffactor_spatial": 16, "ffactor_temporal": 4, "downsample_match_channel": True, "upsample_match_channel": True, "refiner_vae": False} + self.latent_channels = ddconfig['z_channels'] = sd["decoder.conv_in.weight"].shape[1] + self.working_dtypes = [torch.float16, torch.bfloat16, torch.float32] + self.upscale_ratio = (lambda a: max(0, a * 4 - 3), 16, 16) + self.upscale_index_formula = (4, 16, 16) + self.downscale_ratio = (lambda a: max(0, math.floor((a + 3) / 4)), 16, 16) + self.downscale_index_formula = (4, 16, 16) + self.latent_dim = 3 + self.not_video = True + self.first_stage_model = AutoencodingEngine(regularizer_config={'target': "comfy.ldm.models.autoencoder.DiagonalGaussianRegularizer"}, + encoder_config={'target': "comfy.ldm.hunyuan_video.vae_refiner.Encoder", 'params': ddconfig}, + decoder_config={'target': "comfy.ldm.hunyuan_video.vae_refiner.Decoder", 'params': ddconfig}) - self.latent_channels = ddconfig['z_channels'] = sd["decoder.conv_in.weight"].shape[1] - if 'post_quant_conv.weight' in sd: - self.first_stage_model = AutoencoderKL(ddconfig=ddconfig, embed_dim=sd['post_quant_conv.weight'].shape[1]) + self.memory_used_encode = lambda shape, dtype: (2800 * shape[-2] * shape[-1]) * model_management.dtype_size(dtype) + self.memory_used_decode = lambda shape, dtype: (2800 * shape[-3] * shape[-2] * shape[-1] * 16 * 16) * model_management.dtype_size(dtype) else: - self.first_stage_model = AutoencodingEngine(regularizer_config={'target': "comfy.ldm.models.autoencoder.DiagonalGaussianRegularizer"}, - encoder_config={'target': "comfy.ldm.modules.diffusionmodules.model.Encoder", 'params': ddconfig}, - decoder_config={'target': "comfy.ldm.modules.diffusionmodules.model.Decoder", 'params': ddconfig}) + #default SD1.x/SD2.x VAE parameters + ddconfig = {'double_z': True, 'z_channels': 4, 'resolution': 256, 'in_channels': 3, 'out_ch': 3, 'ch': 128, 'ch_mult': [1, 2, 4, 4], 'num_res_blocks': 2, 'attn_resolutions': [], 'dropout': 0.0} + + if 'encoder.down.2.downsample.conv.weight' not in sd and 'decoder.up.3.upsample.conv.weight' not in sd: #Stable diffusion x4 upscaler VAE + ddconfig['ch_mult'] = [1, 2, 4] + self.downscale_ratio = 4 + self.upscale_ratio = 4 + + self.latent_channels = ddconfig['z_channels'] = sd["decoder.conv_in.weight"].shape[1] + if 'post_quant_conv.weight' in sd: + self.first_stage_model = AutoencoderKL(ddconfig=ddconfig, embed_dim=sd['post_quant_conv.weight'].shape[1]) + else: + self.first_stage_model = AutoencodingEngine(regularizer_config={'target': "comfy.ldm.models.autoencoder.DiagonalGaussianRegularizer"}, + encoder_config={'target': "comfy.ldm.modules.diffusionmodules.model.Encoder", 'params': ddconfig}, + decoder_config={'target': "comfy.ldm.modules.diffusionmodules.model.Decoder", 'params': ddconfig}) elif "decoder.layers.1.layers.0.beta" in sd: self.first_stage_model = AudioOobleckVAE() self.memory_used_encode = lambda shape, dtype: (1000 * shape[2]) * model_management.dtype_size(dtype) From bb32d4ec3141333df26fcdaee0c3c08e41b7b249 Mon Sep 17 00:00:00 2001 From: Koratahiu Date: Thu, 2 Oct 2025 00:59:07 +0300 Subject: [PATCH 10/37] feat: Add Epsilon Scaling node for exposure bias correction (#10132) --- comfy_extras/nodes_eps.py | 60 +++++++++++++++++++++++++++++++++++++++ nodes.py | 1 + 2 files changed, 61 insertions(+) create mode 100644 comfy_extras/nodes_eps.py diff --git a/comfy_extras/nodes_eps.py b/comfy_extras/nodes_eps.py new file mode 100644 index 000000000000..c8818f096d9b --- /dev/null +++ b/comfy_extras/nodes_eps.py @@ -0,0 +1,60 @@ +class EpsilonScaling: + """ + Implements the Epsilon Scaling method from 'Elucidating the Exposure Bias in Diffusion Models' + (https://arxiv.org/abs/2308.15321v6). + + This method mitigates exposure bias by scaling the predicted noise during sampling, + which can significantly improve sample quality. This implementation uses the "uniform schedule" + recommended by the paper for its practicality and effectiveness. + """ + @classmethod + def INPUT_TYPES(s): + return { + "required": { + "model": ("MODEL",), + "scaling_factor": ("FLOAT", { + "default": 1.005, + "min": 0.5, + "max": 1.5, + "step": 0.001, + "display": "number" + }), + } + } + + RETURN_TYPES = ("MODEL",) + FUNCTION = "patch" + + CATEGORY = "model_patches/unet" + + def patch(self, model, scaling_factor): + # Prevent division by zero, though the UI's min value should prevent this. + if scaling_factor == 0: + scaling_factor = 1e-9 + + def epsilon_scaling_function(args): + """ + This function is applied after the CFG guidance has been calculated. + It recalculates the denoised latent by scaling the predicted noise. + """ + denoised = args["denoised"] + x = args["input"] + + noise_pred = x - denoised + + scaled_noise_pred = noise_pred / scaling_factor + + new_denoised = x - scaled_noise_pred + + return new_denoised + + # Clone the model patcher to avoid modifying the original model in place + model_clone = model.clone() + + model_clone.set_model_sampler_post_cfg_function(epsilon_scaling_function) + + return (model_clone,) + +NODE_CLASS_MAPPINGS = { + "Epsilon Scaling": EpsilonScaling +} diff --git a/nodes.py b/nodes.py index 1a6784b686bc..88d7129938e5 100644 --- a/nodes.py +++ b/nodes.py @@ -2297,6 +2297,7 @@ async def init_builtin_extra_nodes(): "nodes_gits.py", "nodes_controlnet.py", "nodes_hunyuan.py", + "nodes_eps.py", "nodes_flux.py", "nodes_lora_extract.py", "nodes_torch_compile.py", From 911331c06c16aa80633c5438c58edb32dbfdff50 Mon Sep 17 00:00:00 2001 From: rattus128 <46076784+rattus128@users.noreply.github.com> Date: Thu, 2 Oct 2025 08:40:28 +1000 Subject: [PATCH 11/37] sd: fix VAE tiled fallback VRAM leak (#10139) When the VAE catches this VRAM OOM, it launches the fallback logic straight from the exception context. Python however refs the entire call stack that caused the exception including any local variables for the sake of exception report and debugging. In the case of tensors, this can hold on the references to GBs of VRAM and inhibit the VRAM allocated from freeing them. So dump the except context completely before going back to the VAE via the tiler by getting out of the except block with nothing but a flag. The greately increases the reliability of the tiler fallback, especially on low VRAM cards, as with the bug, if the leak randomly leaked more than the headroom needed for a single tile, the tiler would fallback would OOM and fail the flow. --- comfy/sd.py | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) diff --git a/comfy/sd.py b/comfy/sd.py index 873ad20f2b37..be225ad0357c 100644 --- a/comfy/sd.py +++ b/comfy/sd.py @@ -652,6 +652,7 @@ def encode_tiled_3d(self, samples, tile_t=9999, tile_x=512, tile_y=512, overlap= def decode(self, samples_in, vae_options={}): self.throw_exception_if_invalid() pixel_samples = None + do_tile = False try: memory_used = self.memory_used_decode(samples_in.shape, self.vae_dtype) model_management.load_models_gpu([self.patcher], memory_required=memory_used, force_full_load=self.disable_offload) @@ -667,6 +668,13 @@ def decode(self, samples_in, vae_options={}): pixel_samples[x:x+batch_number] = out except model_management.OOM_EXCEPTION: logging.warning("Warning: Ran out of memory when regular VAE decoding, retrying with tiled VAE decoding.") + #NOTE: We don't know what tensors were allocated to stack variables at the time of the + #exception and the exception itself refs them all until we get out of this except block. + #So we just set a flag for tiler fallback so that tensor gc can happen once the + #exception is fully off the books. + do_tile = True + + if do_tile: dims = samples_in.ndim - 2 if dims == 1 or self.extra_1d_channel is not None: pixel_samples = self.decode_tiled_1d(samples_in) @@ -713,6 +721,7 @@ def encode(self, pixel_samples): self.throw_exception_if_invalid() pixel_samples = self.vae_encode_crop_pixels(pixel_samples) pixel_samples = pixel_samples.movedim(-1, 1) + do_tile = False if self.latent_dim == 3 and pixel_samples.ndim < 5: if not self.not_video: pixel_samples = pixel_samples.movedim(1, 0).unsqueeze(0) @@ -734,6 +743,13 @@ def encode(self, pixel_samples): except model_management.OOM_EXCEPTION: logging.warning("Warning: Ran out of memory when regular VAE encoding, retrying with tiled VAE encoding.") + #NOTE: We don't know what tensors were allocated to stack variables at the time of the + #exception and the exception itself refs them all until we get out of this except block. + #So we just set a flag for tiler fallback so that tensor gc can happen once the + #exception is fully off the books. + do_tile = True + + if do_tile: if self.latent_dim == 3: tile = 256 overlap = tile // 4 From 4965c0e2acf39d84e82cb63dd6cc4400299d0a61 Mon Sep 17 00:00:00 2001 From: rattus128 <46076784+rattus128@users.noreply.github.com> Date: Thu, 2 Oct 2025 08:42:16 +1000 Subject: [PATCH 12/37] WAN: Fix cache VRAM leak on error (#10141) If this suffers an exception (such as a VRAM oom) it will leave the encode() and decode() methods which skips the cleanup of the WAN feature cache. The comfy node cache then ultimately keeps a reference this object which is in turn reffing large tensors from the failed execution. The feature cache is currently setup at a class variable on the encoder/decoder however, the encode and decode functions always clear it on both entry and exit of normal execution. Its likely the design intent is this is usable as a streaming encoder where the input comes in batches, however the functions as they are today don't support that. So simplify by bringing the cache back to local variable, so that if it does VRAM OOM the cache itself is properly garbage when the encode()/decode() functions dissappear from the stack. --- comfy/ldm/wan/vae.py | 37 ++++++++++++++----------------------- 1 file changed, 14 insertions(+), 23 deletions(-) diff --git a/comfy/ldm/wan/vae.py b/comfy/ldm/wan/vae.py index 791596938a26..ccbb258226ba 100644 --- a/comfy/ldm/wan/vae.py +++ b/comfy/ldm/wan/vae.py @@ -468,55 +468,46 @@ def __init__(self, attn_scales, self.temperal_upsample, dropout) def encode(self, x): - self.clear_cache() + conv_idx = [0] + feat_map = [None] * count_conv3d(self.decoder) ## cache t = x.shape[2] iter_ = 1 + (t - 1) // 4 ## 对encode输入的x，按时间拆分为1、4、4、4.... for i in range(iter_): - self._enc_conv_idx = [0] + conv_idx = [0] if i == 0: out = self.encoder( x[:, :, :1, :, :], - feat_cache=self._enc_feat_map, - feat_idx=self._enc_conv_idx) + feat_cache=feat_map, + feat_idx=conv_idx) else: out_ = self.encoder( x[:, :, 1 + 4 * (i - 1):1 + 4 * i, :, :], - feat_cache=self._enc_feat_map, - feat_idx=self._enc_conv_idx) + feat_cache=feat_map, + feat_idx=conv_idx) out = torch.cat([out, out_], 2) mu, log_var = self.conv1(out).chunk(2, dim=1) - self.clear_cache() return mu def decode(self, z): - self.clear_cache() + conv_idx = [0] + feat_map = [None] * count_conv3d(self.decoder) # z: [b,c,t,h,w] iter_ = z.shape[2] x = self.conv2(z) for i in range(iter_): - self._conv_idx = [0] + conv_idx = [0] if i == 0: out = self.decoder( x[:, :, i:i + 1, :, :], - feat_cache=self._feat_map, - feat_idx=self._conv_idx) + feat_cache=feat_map, + feat_idx=conv_idx) else: out_ = self.decoder( x[:, :, i:i + 1, :, :], - feat_cache=self._feat_map, - feat_idx=self._conv_idx) + feat_cache=feat_map, + feat_idx=conv_idx) out = torch.cat([out, out_], 2) - self.clear_cache() return out - - def clear_cache(self): - self._conv_num = count_conv3d(self.decoder) - self._conv_idx = [0] - self._feat_map = [None] * self._conv_num - #cache encode - self._enc_conv_num = count_conv3d(self.encoder) - self._enc_conv_idx = [0] - self._enc_feat_map = [None] * self._enc_conv_num From 0e9d1724be327c79ba86159d868f0b57adb8c384 Mon Sep 17 00:00:00 2001 From: comfyanonymous <121283862+comfyanonymous@users.noreply.github.com> Date: Wed, 1 Oct 2025 21:33:05 -0700 Subject: [PATCH 13/37] Add a .bat to the AMD portable to disable smart memory. (#10153) --- .ci/windows_amd_base_files/README_VERY_IMPORTANT.txt | 5 ++++- .../run_amd_gpu_disable_smart_memory.bat | 2 ++ 2 files changed, 6 insertions(+), 1 deletion(-) create mode 100755 .ci/windows_amd_base_files/run_amd_gpu_disable_smart_memory.bat diff --git a/.ci/windows_amd_base_files/README_VERY_IMPORTANT.txt b/.ci/windows_amd_base_files/README_VERY_IMPORTANT.txt index 570ac33982cc..96a500be2479 100755 --- a/.ci/windows_amd_base_files/README_VERY_IMPORTANT.txt +++ b/.ci/windows_amd_base_files/README_VERY_IMPORTANT.txt @@ -3,10 +3,13 @@ https://www.amd.com/en/resources/support-articles/release-notes/RN-AMDGPU-WINDOW HOW TO RUN: -if you have a AMD gpu: +If you have a AMD gpu: run_amd_gpu.bat +If you have memory issues you can try disabling the smart memory management by running comfyui with: + +run_amd_gpu_disable_smart_memory.bat IF YOU GET A RED ERROR IN THE UI MAKE SURE YOU HAVE A MODEL/CHECKPOINT IN: ComfyUI\models\checkpoints diff --git a/.ci/windows_amd_base_files/run_amd_gpu_disable_smart_memory.bat b/.ci/windows_amd_base_files/run_amd_gpu_disable_smart_memory.bat new file mode 100755 index 000000000000..cece0aeb20b5 --- /dev/null +++ b/.ci/windows_amd_base_files/run_amd_gpu_disable_smart_memory.bat @@ -0,0 +1,2 @@ +.\python_embeded\python.exe -s ComfyUI\main.py --windows-standalone-build --disable-smart-memory +pause From 8f4ee9984c0c3864290e4fea81cfea2ba281717d Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Thu, 2 Oct 2025 23:53:00 +0300 Subject: [PATCH 14/37] convert nodes_morphology.py to V3 schema (#10159) --- comfy_extras/nodes_morphology.py | 116 +++++++++++++++++++------------ 1 file changed, 70 insertions(+), 46 deletions(-) diff --git a/comfy_extras/nodes_morphology.py b/comfy_extras/nodes_morphology.py index 075b26c4024b..67377e1bc94d 100644 --- a/comfy_extras/nodes_morphology.py +++ b/comfy_extras/nodes_morphology.py @@ -1,24 +1,34 @@ import torch import comfy.model_management +from typing_extensions import override +from comfy_api.latest import ComfyExtension, io from kornia.morphology import dilation, erosion, opening, closing, gradient, top_hat, bottom_hat import kornia.color -class Morphology: +class Morphology(io.ComfyNode): @classmethod - def INPUT_TYPES(s): - return {"required": {"image": ("IMAGE",), - "operation": (["erode", "dilate", "open", "close", "gradient", "bottom_hat", "top_hat"],), - "kernel_size": ("INT", {"default": 3, "min": 3, "max": 999, "step": 1}), - }} + def define_schema(cls): + return io.Schema( + node_id="Morphology", + display_name="ImageMorphology", + category="image/postprocessing", + inputs=[ + io.Image.Input("image"), + io.Combo.Input( + "operation", + options=["erode", "dilate", "open", "close", "gradient", "bottom_hat", "top_hat"], + ), + io.Int.Input("kernel_size", default=3, min=3, max=999, step=1), + ], + outputs=[ + io.Image.Output(), + ], + ) - RETURN_TYPES = ("IMAGE",) - FUNCTION = "process" - - CATEGORY = "image/postprocessing" - - def process(self, image, operation, kernel_size): + @classmethod + def execute(cls, image, operation, kernel_size) -> io.NodeOutput: device = comfy.model_management.get_torch_device() kernel = torch.ones(kernel_size, kernel_size, device=device) image_k = image.to(device).movedim(-1, 1) @@ -39,49 +49,63 @@ def process(self, image, operation, kernel_size): else: raise ValueError(f"Invalid operation {operation} for morphology. Must be one of 'erode', 'dilate', 'open', 'close', 'gradient', 'tophat', 'bottomhat'") img_out = output.to(comfy.model_management.intermediate_device()).movedim(1, -1) - return (img_out,) + return io.NodeOutput(img_out) -class ImageRGBToYUV: +class ImageRGBToYUV(io.ComfyNode): @classmethod - def INPUT_TYPES(s): - return {"required": { "image": ("IMAGE",), - }} + def define_schema(cls): + return io.Schema( + node_id="ImageRGBToYUV", + category="image/batch", + inputs=[ + io.Image.Input("image"), + ], + outputs=[ + io.Image.Output(display_name="Y"), + io.Image.Output(display_name="U"), + io.Image.Output(display_name="V"), + ], + ) - RETURN_TYPES = ("IMAGE", "IMAGE", "IMAGE") - RETURN_NAMES = ("Y", "U", "V") - FUNCTION = "execute" - - CATEGORY = "image/batch" - - def execute(self, image): + @classmethod + def execute(cls, image) -> io.NodeOutput: out = kornia.color.rgb_to_ycbcr(image.movedim(-1, 1)).movedim(1, -1) - return (out[..., 0:1].expand_as(image), out[..., 1:2].expand_as(image), out[..., 2:3].expand_as(image)) + return io.NodeOutput(out[..., 0:1].expand_as(image), out[..., 1:2].expand_as(image), out[..., 2:3].expand_as(image)) -class ImageYUVToRGB: +class ImageYUVToRGB(io.ComfyNode): @classmethod - def INPUT_TYPES(s): - return {"required": {"Y": ("IMAGE",), - "U": ("IMAGE",), - "V": ("IMAGE",), - }} + def define_schema(cls): + return io.Schema( + node_id="ImageYUVToRGB", + category="image/batch", + inputs=[ + io.Image.Input("Y"), + io.Image.Input("U"), + io.Image.Input("V"), + ], + outputs=[ + io.Image.Output(), + ], + ) - RETURN_TYPES = ("IMAGE",) - FUNCTION = "execute" - - CATEGORY = "image/batch" - - def execute(self, Y, U, V): + @classmethod + def execute(cls, Y, U, V) -> io.NodeOutput: image = torch.cat([torch.mean(Y, dim=-1, keepdim=True), torch.mean(U, dim=-1, keepdim=True), torch.mean(V, dim=-1, keepdim=True)], dim=-1) out = kornia.color.ycbcr_to_rgb(image.movedim(-1, 1)).movedim(1, -1) - return (out,) + return io.NodeOutput(out) + + +class MorphologyExtension(ComfyExtension): + @override + async def get_node_list(self) -> list[type[io.ComfyNode]]: + return [ + Morphology, + ImageRGBToYUV, + ImageYUVToRGB, + ] + -NODE_CLASS_MAPPINGS = { - "Morphology": Morphology, - "ImageRGBToYUV": ImageRGBToYUV, - "ImageYUVToRGB": ImageYUVToRGB, -} +async def comfy_entrypoint() -> MorphologyExtension: + return MorphologyExtension() -NODE_DISPLAY_NAME_MAPPINGS = { - "Morphology": "ImageMorphology", -} From f6e3e9a456127a7e539929f42ea6cac838197879 Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Fri, 3 Oct 2025 00:50:31 +0300 Subject: [PATCH 15/37] fix(api-nodes): made logging path to be smaller (#10156) --- comfy_api_nodes/apis/client.py | 5 +- comfy_api_nodes/apis/request_logger.py | 72 ++++++++++++++++++++------ 2 files changed, 59 insertions(+), 18 deletions(-) diff --git a/comfy_api_nodes/apis/client.py b/comfy_api_nodes/apis/client.py index 0aed906fbf5b..18a694675e1c 100644 --- a/comfy_api_nodes/apis/client.py +++ b/comfy_api_nodes/apis/client.py @@ -95,6 +95,7 @@ import asyncio import logging import io +import os import socket from aiohttp.client_exceptions import ClientError, ClientResponseError from typing import Dict, Type, Optional, Any, TypeVar, Generic, Callable, Tuple @@ -499,7 +500,9 @@ async def upload_file( else: raise ValueError("File must be BytesIO or str path") - operation_id = f"upload_{upload_url.split('/')[-1]}_{uuid.uuid4().hex[:8]}" + parsed = urlparse(upload_url) + basename = os.path.basename(parsed.path) or parsed.netloc or "upload" + operation_id = f"upload_{basename}_{uuid.uuid4().hex[:8]}" request_logger.log_request_response( operation_id=operation_id, request_method="PUT", diff --git a/comfy_api_nodes/apis/request_logger.py b/comfy_api_nodes/apis/request_logger.py index 42901e1413da..2e0ca53802f1 100644 --- a/comfy_api_nodes/apis/request_logger.py +++ b/comfy_api_nodes/apis/request_logger.py @@ -4,16 +4,18 @@ import datetime import json import logging +import re +import hashlib +from typing import Any + import folder_paths # Get the logger instance logger = logging.getLogger(__name__) + def get_log_directory(): - """ - Ensures the API log directory exists within ComfyUI's temp directory - and returns its path. - """ + """Ensures the API log directory exists within ComfyUI's temp directory and returns its path.""" base_temp_dir = folder_paths.get_temp_directory() log_dir = os.path.join(base_temp_dir, "api_logs") try: @@ -24,42 +26,77 @@ def get_log_directory(): return base_temp_dir return log_dir -def _format_data_for_logging(data): + +def _sanitize_filename_component(name: str) -> str: + if not name: + return "log" + sanitized = re.sub(r"[^A-Za-z0-9._-]+", "_", name) # Replace disallowed characters with underscore + sanitized = sanitized.strip(" ._") # Windows: trailing dots or spaces are not allowed + if not sanitized: + sanitized = "log" + return sanitized + + +def _short_hash(*parts: str, length: int = 10) -> str: + return hashlib.sha1(("|".join(parts)).encode("utf-8")).hexdigest()[:length] + + +def _build_log_filepath(log_dir: str, operation_id: str, request_url: str) -> str: + """Build log filepath. We keep it well under common path length limits aiming for <= 240 characters total.""" + timestamp = datetime.datetime.now().strftime("%Y%m%d_%H%M%S_%f") + slug = _sanitize_filename_component(operation_id) # Best-effort human-readable slug from operation_id + h = _short_hash(operation_id or "", request_url or "") # Short hash ties log to the full operation and URL + + # Compute how much room we have for the slug given the directory length + # Keep total path length reasonably below ~260 on Windows. + max_total_path = 240 + prefix = f"{timestamp}_" + suffix = f"_{h}.log" + if not slug: + slug = "op" + max_filename_len = max(60, max_total_path - len(log_dir) - 1) + max_slug_len = max(8, max_filename_len - len(prefix) - len(suffix)) + if len(slug) > max_slug_len: + slug = slug[:max_slug_len].rstrip(" ._-") + return os.path.join(log_dir, f"{prefix}{slug}{suffix}") + + +def _format_data_for_logging(data: Any) -> str: """Helper to format data (dict, str, bytes) for logging.""" if isinstance(data, bytes): try: - return data.decode('utf-8') # Try to decode as text + return data.decode("utf-8") # Try to decode as text except UnicodeDecodeError: return f"[Binary data of length {len(data)} bytes]" elif isinstance(data, (dict, list)): try: return json.dumps(data, indent=2, ensure_ascii=False) except TypeError: - return str(data) # Fallback for non-serializable objects + return str(data) # Fallback for non-serializable objects return str(data) + def log_request_response( operation_id: str, request_method: str, request_url: str, request_headers: dict | None = None, request_params: dict | None = None, - request_data: any = None, + request_data: Any = None, response_status_code: int | None = None, response_headers: dict | None = None, - response_content: any = None, - error_message: str | None = None + response_content: Any = None, + error_message: str | None = None, ): """ Logs API request and response details to a file in the temp/api_logs directory. + Filenames are sanitized and length-limited for cross-platform safety. + If we still fail to write, we fall back to appending into api.log. """ log_dir = get_log_directory() - timestamp = datetime.datetime.now().strftime("%Y%m%d_%H%M%S_%f") - filename = f"{timestamp}_{operation_id.replace('/', '_').replace(':', '_')}.log" - filepath = os.path.join(log_dir, filename) - - log_content = [] + filepath = _build_log_filepath(log_dir, operation_id, request_url) + log_content: list[str] = [] log_content.append(f"Timestamp: {datetime.datetime.now().isoformat()}") log_content.append(f"Operation ID: {operation_id}") log_content.append("-" * 30 + " REQUEST " + "-" * 30) @@ -69,7 +106,7 @@ def log_request_response( log_content.append(f"Headers:\n{_format_data_for_logging(request_headers)}") if request_params: log_content.append(f"Params:\n{_format_data_for_logging(request_params)}") - if request_data: + if request_data is not None: log_content.append(f"Data/Body:\n{_format_data_for_logging(request_data)}") log_content.append("\n" + "-" * 30 + " RESPONSE " + "-" * 30) @@ -77,7 +114,7 @@ def log_request_response( log_content.append(f"Status Code: {response_status_code}") if response_headers: log_content.append(f"Headers:\n{_format_data_for_logging(response_headers)}") - if response_content: + if response_content is not None: log_content.append(f"Content:\n{_format_data_for_logging(response_content)}") if error_message: log_content.append(f"Error:\n{error_message}") @@ -89,6 +126,7 @@ def log_request_response( except Exception as e: logger.error(f"Error writing API log to {filepath}: {e}") + if __name__ == '__main__': # Example usage (for testing the logger directly) logger.setLevel(logging.DEBUG) From e9364ee279f65d0546fea1796c3cd2e0b7e1965f Mon Sep 17 00:00:00 2001 From: comfyanonymous <121283862+comfyanonymous@users.noreply.github.com> Date: Thu, 2 Oct 2025 14:57:15 -0700 Subject: [PATCH 16/37] Turn on TORCH_ROCM_AOTRITON_ENABLE_EXPERIMENTAL by default. (#10168) --- main.py | 1 + 1 file changed, 1 insertion(+) diff --git a/main.py b/main.py index 70696fcc389f..35857dba8a6b 100644 --- a/main.py +++ b/main.py @@ -115,6 +115,7 @@ def execute_script(script_path): os.environ['MIMALLOC_PURGE_DELAY'] = '0' if __name__ == "__main__": + os.environ['TORCH_ROCM_AOTRITON_ENABLE_EXPERIMENTAL'] = '1' if args.default_device is not None: default_dev = args.default_device devices = list(range(32)) From 1395bce9f707e52ec613eeaa87ea690518cfe0a8 Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Fri, 3 Oct 2025 01:20:29 +0300 Subject: [PATCH 17/37] update example_node to use V3 schema (#9723) --- custom_nodes/example_node.py.example | 163 ++++++++++++--------------- 1 file changed, 69 insertions(+), 94 deletions(-) diff --git a/custom_nodes/example_node.py.example b/custom_nodes/example_node.py.example index 29ab2aa72319..779c35787ea8 100644 --- a/custom_nodes/example_node.py.example +++ b/custom_nodes/example_node.py.example @@ -1,96 +1,70 @@ -class Example: +from typing_extensions import override + +from comfy_api.latest import ComfyExtension, io + + +class Example(io.ComfyNode): """ - A example node + An example node Class methods ------------- - INPUT_TYPES (dict): - Tell the main program input parameters of nodes. - IS_CHANGED: + define_schema (io.Schema): + Tell the main program the metadata, input, output parameters of nodes. + fingerprint_inputs: optional method to control when the node is re executed. + check_lazy_status: + optional method to control list of input names that need to be evaluated. - Attributes - ---------- - RETURN_TYPES (`tuple`): - The type of each element in the output tuple. - RETURN_NAMES (`tuple`): - Optional: The name of each output in the output tuple. - FUNCTION (`str`): - The name of the entry-point method. For example, if `FUNCTION = "execute"` then it will run Example().execute() - OUTPUT_NODE ([`bool`]): - If this node is an output node that outputs a result/image from the graph. The SaveImage node is an example. - The backend iterates on these output nodes and tries to execute all their parents if their parent graph is properly connected. - Assumed to be False if not present. - CATEGORY (`str`): - The category the node should appear in the UI. - DEPRECATED (`bool`): - Indicates whether the node is deprecated. Deprecated nodes are hidden by default in the UI, but remain - functional in existing workflows that use them. - EXPERIMENTAL (`bool`): - Indicates whether the node is experimental. Experimental nodes are marked as such in the UI and may be subject to - significant changes or removal in future versions. Use with caution in production workflows. - execute(s) -> tuple || None: - The entry point method. The name of this method must be the same as the value of property `FUNCTION`. - For example, if `FUNCTION = "execute"` then this method's name must be `execute`, if `FUNCTION = "foo"` then it must be `foo`. """ - def __init__(self): - pass @classmethod - def INPUT_TYPES(s): + def define_schema(cls) -> io.Schema: """ - Return a dictionary which contains config for all input fields. - Some types (string): "MODEL", "VAE", "CLIP", "CONDITIONING", "LATENT", "IMAGE", "INT", "STRING", "FLOAT". - Input types "INT", "STRING" or "FLOAT" are special values for fields on the node. - The type can be a list for selection. - - Returns: `dict`: - - Key input_fields_group (`string`): Can be either required, hidden or optional. A node class must have property `required` - - Value input_fields (`dict`): Contains input fields config: - * Key field_name (`string`): Name of a entry-point method's argument - * Value field_config (`tuple`): - + First value is a string indicate the type of field or a list for selection. - + Second value is a config for type "INT", "STRING" or "FLOAT". + Return a schema which contains all information about the node. + Some types: "Model", "Vae", "Clip", "Conditioning", "Latent", "Image", "Int", "String", "Float", "Combo". + For outputs the "io.Model.Output" should be used, for inputs the "io.Model.Input" can be used. + The type can be a "Combo" - this will be a list for selection. """ - return { - "required": { - "image": ("IMAGE",), - "int_field": ("INT", { - "default": 0, - "min": 0, #Minimum value - "max": 4096, #Maximum value - "step": 64, #Slider's step - "display": "number", # Cosmetic only: display as "number" or "slider" - "lazy": True # Will only be evaluated if check_lazy_status requires it - }), - "float_field": ("FLOAT", { - "default": 1.0, - "min": 0.0, - "max": 10.0, - "step": 0.01, - "round": 0.001, #The value representing the precision to round to, will be set to the step value by default. Can be set to False to disable rounding. - "display": "number", - "lazy": True - }), - "print_to_screen": (["enable", "disable"],), - "string_field": ("STRING", { - "multiline": False, #True if you want the field to look like the one on the ClipTextEncode node - "default": "Hello World!", - "lazy": True - }), - }, - } - - RETURN_TYPES = ("IMAGE",) - #RETURN_NAMES = ("image_output_name",) - - FUNCTION = "test" - - #OUTPUT_NODE = False - - CATEGORY = "Example" - - def check_lazy_status(self, image, string_field, int_field, float_field, print_to_screen): + return io.Schema( + node_id="Example", + display_name="Example Node", + category="Example", + inputs=[ + io.Image.Input("image"), + io.Int.Input( + "int_field", + min=0, + max=4096, + step=64, # Slider's step + display_mode=io.NumberDisplay.number, # Cosmetic only: display as "number" or "slider" + lazy=True, # Will only be evaluated if check_lazy_status requires it + ), + io.Float.Input( + "float_field", + default=1.0, + min=0.0, + max=10.0, + step=0.01, + round=0.001, #The value representing the precision to round to, will be set to the step value by default. Can be set to False to disable rounding. + display_mode=io.NumberDisplay.number, + lazy=True, + ), + io.Combo.Input("print_to_screen", options=["enable", "disable"]), + io.String.Input( + "string_field", + multiline=False, # True if you want the field to look like the one on the ClipTextEncode node + default="Hello world!", + lazy=True, + ) + ], + outputs=[ + io.Image.Output(), + ], + ) + + @classmethod + def check_lazy_status(cls, image, string_field, int_field, float_field, print_to_screen): """ Return a list of input names that need to be evaluated. @@ -107,7 +81,8 @@ class Example: else: return [] - def test(self, image, string_field, int_field, float_field, print_to_screen): + @classmethod + def execute(cls, image, string_field, int_field, float_field, print_to_screen) -> io.NodeOutput: if print_to_screen == "enable": print(f"""Your input contains: string_field aka input text: {string_field} @@ -116,7 +91,7 @@ class Example: """) #do some processing on the image, in this example I just invert it image = 1.0 - image - return (image,) + return io.NodeOutput(image) """ The node will always be re executed if any of the inputs change but @@ -127,7 +102,7 @@ class Example: changes between executions the LoadImage node is executed again. """ #@classmethod - #def IS_CHANGED(s, image, string_field, int_field, float_field, print_to_screen): + #def fingerprint_inputs(s, image, string_field, int_field, float_field, print_to_screen): # return "" # Set the web directory, any .js file in that directory will be loaded by the frontend as a frontend extension @@ -143,13 +118,13 @@ async def get_hello(request): return web.json_response("hello") -# A dictionary that contains all nodes you want to export with their names -# NOTE: names should be globally unique -NODE_CLASS_MAPPINGS = { - "Example": Example -} +class ExampleExtension(ComfyExtension): + @override + async def get_node_list(self) -> list[type[io.ComfyNode]]: + return [ + Example, + ] + -# A dictionary that contains the friendly/humanly readable titles for the nodes -NODE_DISPLAY_NAME_MAPPINGS = { - "Example": "Example Node" -} +async def comfy_entrypoint() -> ExampleExtension: # ComfyUI calls this to load your extension and its nodes. + return ExampleExtension() From 4ffea0e864275301329ddb5ecc3fbc7211d7a802 Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Fri, 3 Oct 2025 02:14:28 +0300 Subject: [PATCH 18/37] feat(linter, api-nodes): add pylint for comfy_api_nodes folder (#10157) --- .github/workflows/ruff.yml | 25 ++++++++++++++ comfy_api_nodes/apis/__init__.py | 1 + comfy_api_nodes/apis/client.py | 2 +- comfy_api_nodes/apis/rodin_api.py | 4 --- pyproject.toml | 54 +++++++++++++++++++++++++++++++ 5 files changed, 81 insertions(+), 5 deletions(-) diff --git a/.github/workflows/ruff.yml b/.github/workflows/ruff.yml index 4c1a025948b4..b24d86a6ba55 100644 --- a/.github/workflows/ruff.yml +++ b/.github/workflows/ruff.yml @@ -21,3 +21,28 @@ jobs: - name: Run Ruff run: ruff check . + + pylint: + name: Run Pylint + runs-on: ubuntu-latest + + steps: + - name: Checkout repository + uses: actions/checkout@v4 + + - name: Set up Python + uses: actions/setup-python@v4 + with: + python-version: '3.12' + + - name: Install requirements + run: | + python -m pip install --upgrade pip + pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu + pip install -r requirements.txt + + - name: Install Pylint + run: pip install pylint + + - name: Run Pylint + run: pylint comfy_api_nodes diff --git a/comfy_api_nodes/apis/__init__.py b/comfy_api_nodes/apis/__init__.py index 78a23db30c6e..98f9e540d0ea 100644 --- a/comfy_api_nodes/apis/__init__.py +++ b/comfy_api_nodes/apis/__init__.py @@ -2,6 +2,7 @@ # filename: filtered-openapi.yaml # timestamp: 2025-07-30T08:54:00+00:00 +# pylint: disable from __future__ import annotations from datetime import date, datetime diff --git a/comfy_api_nodes/apis/client.py b/comfy_api_nodes/apis/client.py index 18a694675e1c..79de3c262450 100644 --- a/comfy_api_nodes/apis/client.py +++ b/comfy_api_nodes/apis/client.py @@ -535,7 +535,7 @@ async def upload_file( request_method="PUT", request_url=upload_url, response_status_code=e.status if hasattr(e, "status") else None, - response_headers=dict(e.headers) if getattr(e, "headers") else None, + response_headers=dict(e.headers) if hasattr(e, "headers") else None, response_content=None, error_message=f"{type(e).__name__}: {str(e)}", ) diff --git a/comfy_api_nodes/apis/rodin_api.py b/comfy_api_nodes/apis/rodin_api.py index 02cf42c29cb2..fc26a6e73a0f 100644 --- a/comfy_api_nodes/apis/rodin_api.py +++ b/comfy_api_nodes/apis/rodin_api.py @@ -52,7 +52,3 @@ class RodinResourceItem(BaseModel): class Rodin3DDownloadResponse(BaseModel): list: List[RodinResourceItem] = Field(..., description="Source List") - - - - diff --git a/pyproject.toml b/pyproject.toml index d0a76c6d06ee..598af4157273 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -22,3 +22,57 @@ lint.select = [ "F", ] exclude = ["*.ipynb", "**/generated/*.pyi"] + +[tool.pylint] +master.py-version = "3.9" +master.extension-pkg-allow-list = [ + "pydantic", +] +reports.output-format = "colorized" +similarities.ignore-imports = "yes" +messages_control.disable = [ + "missing-module-docstring", + "missing-class-docstring", + "missing-function-docstring", + "line-too-long", + "too-few-public-methods", + "too-many-public-methods", + "too-many-instance-attributes", + "too-many-positional-arguments", + "broad-exception-raised", + "too-many-lines", + "invalid-name", + "unused-argument", + "broad-exception-caught", + "consider-using-with", + "fixme", + "too-many-statements", + "too-many-branches", + "too-many-locals", + "too-many-arguments", + "duplicate-code", + "abstract-method", + "superfluous-parens", + "arguments-differ", + "redefined-builtin", + "unnecessary-lambda", + "dangerous-default-value", + # next warnings should be fixed in future + "bad-classmethod-argument", # Class method should have 'cls' as first argument + "wrong-import-order", # Standard imports should be placed before third party imports + "logging-fstring-interpolation", # Use lazy % formatting in logging functions + "ungrouped-imports", + "unnecessary-pass", + "unidiomatic-typecheck", + "unnecessary-lambda-assignment", + "bad-indentation", + "no-else-return", + "no-else-raise", + "invalid-overridden-method", + "unused-variable", + "pointless-string-statement", + "inconsistent-return-statements", + "import-outside-toplevel", + "reimported", + "redefined-outer-name", +] From ed3ca78e080d697b6cf29497c07e14ee9c27a3ac Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Fri, 3 Oct 2025 21:26:34 +0300 Subject: [PATCH 19/37] feat(api-nodes): add kling-2-5-turbo to txt2video and img2video nodes (#10155) --- comfy_api_nodes/apis/__init__.py | 2 ++ comfy_api_nodes/nodes_kling.py | 2 ++ 2 files changed, 4 insertions(+) diff --git a/comfy_api_nodes/apis/__init__.py b/comfy_api_nodes/apis/__init__.py index 98f9e540d0ea..ee2aa1ce643e 100644 --- a/comfy_api_nodes/apis/__init__.py +++ b/comfy_api_nodes/apis/__init__.py @@ -1321,6 +1321,7 @@ class KlingTextToVideoModelName(str, Enum): kling_v1 = 'kling-v1' kling_v1_6 = 'kling-v1-6' kling_v2_1_master = 'kling-v2-1-master' + kling_v2_5_turbo = 'kling-v2-5-turbo' class KlingVideoGenAspectRatio(str, Enum): @@ -1355,6 +1356,7 @@ class KlingVideoGenModelName(str, Enum): kling_v2_master = 'kling-v2-master' kling_v2_1 = 'kling-v2-1' kling_v2_1_master = 'kling-v2-1-master' + kling_v2_5_turbo = 'kling-v2-5-turbo' class KlingVideoResult(BaseModel): diff --git a/comfy_api_nodes/nodes_kling.py b/comfy_api_nodes/nodes_kling.py index 5f55b2cc9b49..d8646f1060ca 100644 --- a/comfy_api_nodes/nodes_kling.py +++ b/comfy_api_nodes/nodes_kling.py @@ -423,6 +423,8 @@ def get_mode_string_mapping() -> dict[str, tuple[str, str, str]]: "standard mode / 10s duration / kling-v2-master": ("std", "10", "kling-v2-master"), "pro mode / 5s duration / kling-v2-1-master": ("pro", "5", "kling-v2-1-master"), "pro mode / 10s duration / kling-v2-1-master": ("pro", "10", "kling-v2-1-master"), + "pro mode / 5s duration / kling-v2-5-turbo": ("pro", "5", "kling-v2-5-turbo"), + "pro mode / 10s duration / kling-v2-5-turbo": ("pro", "10", "kling-v2-5-turbo"), } @classmethod From 8a293372ecdea0ff8647921eaf3bb10c3d992abf Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Fri, 3 Oct 2025 21:40:27 +0300 Subject: [PATCH 20/37] fix(api-nodes): reimport of base64 in Gemini node (#10181) --- comfy_api_nodes/nodes_gemini.py | 1 - pyproject.toml | 1 - 2 files changed, 2 deletions(-) diff --git a/comfy_api_nodes/nodes_gemini.py b/comfy_api_nodes/nodes_gemini.py index baa379b75448..151cb404465b 100644 --- a/comfy_api_nodes/nodes_gemini.py +++ b/comfy_api_nodes/nodes_gemini.py @@ -490,7 +490,6 @@ def create_file_part(self, file_path: str) -> GeminiPart: # Use base64 string directly, not the data URI with open(file_path, "rb") as f: file_content = f.read() - import base64 base64_str = base64.b64encode(file_content).decode("utf-8") return GeminiPart( diff --git a/pyproject.toml b/pyproject.toml index 598af4157273..7952f7f37295 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -73,6 +73,5 @@ messages_control.disable = [ "pointless-string-statement", "inconsistent-return-statements", "import-outside-toplevel", - "reimported", "redefined-outer-name", ] From c2c5a7d5f80579bb44c11de0ce6eff94d1c111b9 Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Fri, 3 Oct 2025 21:41:06 +0300 Subject: [PATCH 21/37] fix(api-nodes): bad indentation in Recraft API node function (#10175) --- comfy_api_nodes/nodes_recraft.py | 82 ++++++++++++++++---------------- pyproject.toml | 1 - 2 files changed, 41 insertions(+), 42 deletions(-) diff --git a/comfy_api_nodes/nodes_recraft.py b/comfy_api_nodes/nodes_recraft.py index c8516b368208..a006104b7350 100644 --- a/comfy_api_nodes/nodes_recraft.py +++ b/comfy_api_nodes/nodes_recraft.py @@ -38,48 +38,48 @@ async def handle_recraft_file_request( - image: torch.Tensor, - path: str, - mask: torch.Tensor=None, - total_pixels=4096*4096, - timeout=1024, - request=None, - auth_kwargs: dict[str,str] = None, - ) -> list[BytesIO]: - """ - Handle sending common Recraft file-only request to get back file bytes. - """ - if request is None: - request = EmptyRequest() - - files = { - 'image': tensor_to_bytesio(image, total_pixels=total_pixels).read() - } - if mask is not None: - files['mask'] = tensor_to_bytesio(mask, total_pixels=total_pixels).read() - - operation = SynchronousOperation( - endpoint=ApiEndpoint( - path=path, - method=HttpMethod.POST, - request_model=type(request), - response_model=RecraftImageGenerationResponse, - ), - request=request, - files=files, - content_type="multipart/form-data", - auth_kwargs=auth_kwargs, - multipart_parser=recraft_multipart_parser, - ) - response: RecraftImageGenerationResponse = await operation.execute() - all_bytesio = [] - if response.image is not None: - all_bytesio.append(await download_url_to_bytesio(response.image.url, timeout=timeout)) - else: - for data in response.data: - all_bytesio.append(await download_url_to_bytesio(data.url, timeout=timeout)) + image: torch.Tensor, + path: str, + mask: torch.Tensor=None, + total_pixels=4096*4096, + timeout=1024, + request=None, + auth_kwargs: dict[str,str] = None, +) -> list[BytesIO]: + """ + Handle sending common Recraft file-only request to get back file bytes. + """ + if request is None: + request = EmptyRequest() + + files = { + 'image': tensor_to_bytesio(image, total_pixels=total_pixels).read() + } + if mask is not None: + files['mask'] = tensor_to_bytesio(mask, total_pixels=total_pixels).read() + + operation = SynchronousOperation( + endpoint=ApiEndpoint( + path=path, + method=HttpMethod.POST, + request_model=type(request), + response_model=RecraftImageGenerationResponse, + ), + request=request, + files=files, + content_type="multipart/form-data", + auth_kwargs=auth_kwargs, + multipart_parser=recraft_multipart_parser, + ) + response: RecraftImageGenerationResponse = await operation.execute() + all_bytesio = [] + if response.image is not None: + all_bytesio.append(await download_url_to_bytesio(response.image.url, timeout=timeout)) + else: + for data in response.data: + all_bytesio.append(await download_url_to_bytesio(data.url, timeout=timeout)) - return all_bytesio + return all_bytesio def recraft_multipart_parser(data, parent_key=None, formatter: callable=None, converted_to_check: list[list]=None, is_list=False) -> dict: diff --git a/pyproject.toml b/pyproject.toml index 7952f7f37295..240919a4390e 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -65,7 +65,6 @@ messages_control.disable = [ "unnecessary-pass", "unidiomatic-typecheck", "unnecessary-lambda-assignment", - "bad-indentation", "no-else-return", "no-else-raise", "invalid-overridden-method", From 3e68bc342cd60b909b4117c1b68a3afc62ef875c Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Fri, 3 Oct 2025 21:43:54 +0300 Subject: [PATCH 22/37] convert nodes_torch_compile.py to V3 schema (#10173) --- comfy_extras/nodes_torch_compile.py | 46 +++++++++++++++++++---------- 1 file changed, 31 insertions(+), 15 deletions(-) diff --git a/comfy_extras/nodes_torch_compile.py b/comfy_extras/nodes_torch_compile.py index 6055366784d0..adbeece2f384 100644 --- a/comfy_extras/nodes_torch_compile.py +++ b/comfy_extras/nodes_torch_compile.py @@ -1,23 +1,39 @@ +from typing_extensions import override +from comfy_api.latest import ComfyExtension, io from comfy_api.torch_helpers import set_torch_compile_wrapper -class TorchCompileModel: +class TorchCompileModel(io.ComfyNode): @classmethod - def INPUT_TYPES(s): - return {"required": { "model": ("MODEL",), - "backend": (["inductor", "cudagraphs"],), - }} - RETURN_TYPES = ("MODEL",) - FUNCTION = "patch" + def define_schema(cls) -> io.Schema: + return io.Schema( + node_id="TorchCompileModel", + category="_for_testing", + inputs=[ + io.Model.Input("model"), + io.Combo.Input( + "backend", + options=["inductor", "cudagraphs"], + ), + ], + outputs=[io.Model.Output()], + is_experimental=True, + ) - CATEGORY = "_for_testing" - EXPERIMENTAL = True - - def patch(self, model, backend): + @classmethod + def execute(cls, model, backend) -> io.NodeOutput: m = model.clone() set_torch_compile_wrapper(model=m, backend=backend) - return (m, ) + return io.NodeOutput(m) + + +class TorchCompileExtension(ComfyExtension): + @override + async def get_node_list(self) -> list[type[io.ComfyNode]]: + return [ + TorchCompileModel, + ] + -NODE_CLASS_MAPPINGS = { - "TorchCompileModel": TorchCompileModel, -} +async def comfy_entrypoint() -> TorchCompileExtension: + return TorchCompileExtension() From d7aa414141f02a456801704a3da323fa2ed8f5cc Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Fri, 3 Oct 2025 21:45:02 +0300 Subject: [PATCH 23/37] convert nodes_eps.py to V3 schema (#10172) --- comfy_extras/nodes_eps.py | 64 ++++++++++++++++++++++++--------------- 1 file changed, 39 insertions(+), 25 deletions(-) diff --git a/comfy_extras/nodes_eps.py b/comfy_extras/nodes_eps.py index c8818f096d9b..7852d85e5aca 100644 --- a/comfy_extras/nodes_eps.py +++ b/comfy_extras/nodes_eps.py @@ -1,4 +1,9 @@ -class EpsilonScaling: +from typing_extensions import override + +from comfy_api.latest import ComfyExtension, io + + +class EpsilonScaling(io.ComfyNode): """ Implements the Epsilon Scaling method from 'Elucidating the Exposure Bias in Diffusion Models' (https://arxiv.org/abs/2308.15321v6). @@ -8,26 +13,28 @@ class EpsilonScaling: recommended by the paper for its practicality and effectiveness. """ @classmethod - def INPUT_TYPES(s): - return { - "required": { - "model": ("MODEL",), - "scaling_factor": ("FLOAT", { - "default": 1.005, - "min": 0.5, - "max": 1.5, - "step": 0.001, - "display": "number" - }), - } - } - - RETURN_TYPES = ("MODEL",) - FUNCTION = "patch" - - CATEGORY = "model_patches/unet" - - def patch(self, model, scaling_factor): + def define_schema(cls): + return io.Schema( + node_id="Epsilon Scaling", + category="model_patches/unet", + inputs=[ + io.Model.Input("model"), + io.Float.Input( + "scaling_factor", + default=1.005, + min=0.5, + max=1.5, + step=0.001, + display_mode=io.NumberDisplay.number, + ), + ], + outputs=[ + io.Model.Output(), + ], + ) + + @classmethod + def execute(cls, model, scaling_factor) -> io.NodeOutput: # Prevent division by zero, though the UI's min value should prevent this. if scaling_factor == 0: scaling_factor = 1e-9 @@ -53,8 +60,15 @@ def epsilon_scaling_function(args): model_clone.set_model_sampler_post_cfg_function(epsilon_scaling_function) - return (model_clone,) + return io.NodeOutput(model_clone) + + +class EpsilonScalingExtension(ComfyExtension): + @override + async def get_node_list(self) -> list[type[io.ComfyNode]]: + return [ + EpsilonScaling, + ] -NODE_CLASS_MAPPINGS = { - "Epsilon Scaling": EpsilonScaling -} +async def comfy_entrypoint() -> EpsilonScalingExtension: + return EpsilonScalingExtension() From 8c26d7bbe6663f589f0a9562921aafb3c48955c6 Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Fri, 3 Oct 2025 21:48:21 +0300 Subject: [PATCH 24/37] convert nodes_pixverse.py to V3 schema (#10177) --- comfy_api_nodes/nodes_pixverse.py | 449 +++++++++++++++--------------- 1 file changed, 227 insertions(+), 222 deletions(-) diff --git a/comfy_api_nodes/nodes_pixverse.py b/comfy_api_nodes/nodes_pixverse.py index 7c5a52feb42c..eb98e9653f43 100644 --- a/comfy_api_nodes/nodes_pixverse.py +++ b/comfy_api_nodes/nodes_pixverse.py @@ -1,5 +1,7 @@ from inspect import cleandoc from typing import Optional +from typing_extensions import override +from io import BytesIO from comfy_api_nodes.apis.pixverse_api import ( PixverseTextVideoRequest, PixverseImageVideoRequest, @@ -26,12 +28,11 @@ tensor_to_bytesio, validate_string, ) -from comfy.comfy_types.node_typing import IO, ComfyNodeABC from comfy_api.input_impl import VideoFromFile +from comfy_api.latest import ComfyExtension, io as comfy_io import torch import aiohttp -from io import BytesIO AVERAGE_DURATION_T2V = 32 @@ -72,100 +73,101 @@ async def upload_image_to_pixverse(image: torch.Tensor, auth_kwargs=None): return response_upload.Resp.img_id -class PixverseTemplateNode: +class PixverseTemplateNode(comfy_io.ComfyNode): """ Select template for PixVerse Video generation. """ - RETURN_TYPES = (PixverseIO.TEMPLATE,) - RETURN_NAMES = ("pixverse_template",) - FUNCTION = "create_template" - CATEGORY = "api node/video/PixVerse" - @classmethod - def INPUT_TYPES(s): - return { - "required": { - "template": (list(pixverse_templates.keys()),), - } - } + def define_schema(cls) -> comfy_io.Schema: + return comfy_io.Schema( + node_id="PixverseTemplateNode", + display_name="PixVerse Template", + category="api node/video/PixVerse", + inputs=[ + comfy_io.Combo.Input("template", options=[list(pixverse_templates.keys())]), + ], + outputs=[comfy_io.Custom(PixverseIO.TEMPLATE).Output(display_name="pixverse_template")], + ) - def create_template(self, template: str): + @classmethod + def execute(cls, template: str) -> comfy_io.NodeOutput: template_id = pixverse_templates.get(template, None) if template_id is None: raise Exception(f"Template '{template}' is not recognized.") # just return the integer - return (template_id,) + return comfy_io.NodeOutput(template_id) -class PixverseTextToVideoNode(ComfyNodeABC): +class PixverseTextToVideoNode(comfy_io.ComfyNode): """ Generates videos based on prompt and output_size. """ - RETURN_TYPES = (IO.VIDEO,) - DESCRIPTION = cleandoc(__doc__ or "") # Handle potential None value - FUNCTION = "api_call" - API_NODE = True - CATEGORY = "api node/video/PixVerse" - @classmethod - def INPUT_TYPES(s): - return { - "required": { - "prompt": ( - IO.STRING, - { - "multiline": True, - "default": "", - "tooltip": "Prompt for the video generation", - }, + def define_schema(cls) -> comfy_io.Schema: + return comfy_io.Schema( + node_id="PixverseTextToVideoNode", + display_name="PixVerse Text to Video", + category="api node/video/PixVerse", + description=cleandoc(cls.__doc__ or ""), + inputs=[ + comfy_io.String.Input( + "prompt", + multiline=True, + default="", + tooltip="Prompt for the video generation", ), - "aspect_ratio": ([ratio.value for ratio in PixverseAspectRatio],), - "quality": ( - [resolution.value for resolution in PixverseQuality], - { - "default": PixverseQuality.res_540p, - }, + comfy_io.Combo.Input( + "aspect_ratio", + options=[ratio.value for ratio in PixverseAspectRatio], ), - "duration_seconds": ([dur.value for dur in PixverseDuration],), - "motion_mode": ([mode.value for mode in PixverseMotionMode],), - "seed": ( - IO.INT, - { - "default": 0, - "min": 0, - "max": 2147483647, - "control_after_generate": True, - "tooltip": "Seed for video generation.", - }, + comfy_io.Combo.Input( + "quality", + options=[resolution.value for resolution in PixverseQuality], + default=PixverseQuality.res_540p, ), - }, - "optional": { - "negative_prompt": ( - IO.STRING, - { - "default": "", - "forceInput": True, - "tooltip": "An optional text description of undesired elements on an image.", - }, + comfy_io.Combo.Input( + "duration_seconds", + options=[dur.value for dur in PixverseDuration], ), - "pixverse_template": ( - PixverseIO.TEMPLATE, - { - "tooltip": "An optional template to influence style of generation, created by the PixVerse Template node." - }, + comfy_io.Combo.Input( + "motion_mode", + options=[mode.value for mode in PixverseMotionMode], ), - }, - "hidden": { - "auth_token": "AUTH_TOKEN_COMFY_ORG", - "comfy_api_key": "API_KEY_COMFY_ORG", - "unique_id": "UNIQUE_ID", - }, - } + comfy_io.Int.Input( + "seed", + default=0, + min=0, + max=2147483647, + control_after_generate=True, + tooltip="Seed for video generation.", + ), + comfy_io.String.Input( + "negative_prompt", + default="", + force_input=True, + tooltip="An optional text description of undesired elements on an image.", + optional=True, + ), + comfy_io.Custom(PixverseIO.TEMPLATE).Input( + "pixverse_template", + tooltip="An optional template to influence style of generation, created by the PixVerse Template node.", + optional=True, + ), + ], + outputs=[comfy_io.Video.Output()], + hidden=[ + comfy_io.Hidden.auth_token_comfy_org, + comfy_io.Hidden.api_key_comfy_org, + comfy_io.Hidden.unique_id, + ], + is_api_node=True, + ) - async def api_call( - self, + @classmethod + async def execute( + cls, prompt: str, aspect_ratio: str, quality: str, @@ -174,9 +176,7 @@ async def api_call( seed, negative_prompt: str = None, pixverse_template: int = None, - unique_id: Optional[str] = None, - **kwargs, - ): + ) -> comfy_io.NodeOutput: validate_string(prompt, strip_whitespace=False) # 1080p is limited to 5 seconds duration # only normal motion_mode supported for 1080p or for non-5 second duration @@ -186,6 +186,10 @@ async def api_call( elif duration_seconds != PixverseDuration.dur_5: motion_mode = PixverseMotionMode.normal + auth = { + "auth_token": cls.hidden.auth_token_comfy_org, + "comfy_api_key": cls.hidden.api_key_comfy_org, + } operation = SynchronousOperation( endpoint=ApiEndpoint( path="/proxy/pixverse/video/text/generate", @@ -203,7 +207,7 @@ async def api_call( template_id=pixverse_template, seed=seed, ), - auth_kwargs=kwargs, + auth_kwargs=auth, ) response_api = await operation.execute() @@ -224,8 +228,8 @@ async def api_call( PixverseStatus.deleted, ], status_extractor=lambda x: x.Resp.status, - auth_kwargs=kwargs, - node_id=unique_id, + auth_kwargs=auth, + node_id=cls.hidden.unique_id, result_url_extractor=get_video_url_from_response, estimated_duration=AVERAGE_DURATION_T2V, ) @@ -233,77 +237,75 @@ async def api_call( async with aiohttp.ClientSession() as session: async with session.get(response_poll.Resp.url) as vid_response: - return (VideoFromFile(BytesIO(await vid_response.content.read())),) + return comfy_io.NodeOutput(VideoFromFile(BytesIO(await vid_response.content.read()))) -class PixverseImageToVideoNode(ComfyNodeABC): +class PixverseImageToVideoNode(comfy_io.ComfyNode): """ Generates videos based on prompt and output_size. """ - RETURN_TYPES = (IO.VIDEO,) - DESCRIPTION = cleandoc(__doc__ or "") # Handle potential None value - FUNCTION = "api_call" - API_NODE = True - CATEGORY = "api node/video/PixVerse" - @classmethod - def INPUT_TYPES(s): - return { - "required": { - "image": (IO.IMAGE,), - "prompt": ( - IO.STRING, - { - "multiline": True, - "default": "", - "tooltip": "Prompt for the video generation", - }, + def define_schema(cls) -> comfy_io.Schema: + return comfy_io.Schema( + node_id="PixverseImageToVideoNode", + display_name="PixVerse Image to Video", + category="api node/video/PixVerse", + description=cleandoc(cls.__doc__ or ""), + inputs=[ + comfy_io.Image.Input("image"), + comfy_io.String.Input( + "prompt", + multiline=True, + default="", + tooltip="Prompt for the video generation", ), - "quality": ( - [resolution.value for resolution in PixverseQuality], - { - "default": PixverseQuality.res_540p, - }, + comfy_io.Combo.Input( + "quality", + options=[resolution.value for resolution in PixverseQuality], + default=PixverseQuality.res_540p, ), - "duration_seconds": ([dur.value for dur in PixverseDuration],), - "motion_mode": ([mode.value for mode in PixverseMotionMode],), - "seed": ( - IO.INT, - { - "default": 0, - "min": 0, - "max": 2147483647, - "control_after_generate": True, - "tooltip": "Seed for video generation.", - }, + comfy_io.Combo.Input( + "duration_seconds", + options=[dur.value for dur in PixverseDuration], ), - }, - "optional": { - "negative_prompt": ( - IO.STRING, - { - "default": "", - "forceInput": True, - "tooltip": "An optional text description of undesired elements on an image.", - }, + comfy_io.Combo.Input( + "motion_mode", + options=[mode.value for mode in PixverseMotionMode], ), - "pixverse_template": ( - PixverseIO.TEMPLATE, - { - "tooltip": "An optional template to influence style of generation, created by the PixVerse Template node." - }, + comfy_io.Int.Input( + "seed", + default=0, + min=0, + max=2147483647, + control_after_generate=True, + tooltip="Seed for video generation.", ), - }, - "hidden": { - "auth_token": "AUTH_TOKEN_COMFY_ORG", - "comfy_api_key": "API_KEY_COMFY_ORG", - "unique_id": "UNIQUE_ID", - }, - } + comfy_io.String.Input( + "negative_prompt", + default="", + force_input=True, + tooltip="An optional text description of undesired elements on an image.", + optional=True, + ), + comfy_io.Custom(PixverseIO.TEMPLATE).Input( + "pixverse_template", + tooltip="An optional template to influence style of generation, created by the PixVerse Template node.", + optional=True, + ), + ], + outputs=[comfy_io.Video.Output()], + hidden=[ + comfy_io.Hidden.auth_token_comfy_org, + comfy_io.Hidden.api_key_comfy_org, + comfy_io.Hidden.unique_id, + ], + is_api_node=True, + ) - async def api_call( - self, + @classmethod + async def execute( + cls, image: torch.Tensor, prompt: str, quality: str, @@ -312,11 +314,13 @@ async def api_call( seed, negative_prompt: str = None, pixverse_template: int = None, - unique_id: Optional[str] = None, - **kwargs, - ): + ) -> comfy_io.NodeOutput: validate_string(prompt, strip_whitespace=False) - img_id = await upload_image_to_pixverse(image, auth_kwargs=kwargs) + auth = { + "auth_token": cls.hidden.auth_token_comfy_org, + "comfy_api_key": cls.hidden.api_key_comfy_org, + } + img_id = await upload_image_to_pixverse(image, auth_kwargs=auth) # 1080p is limited to 5 seconds duration # only normal motion_mode supported for 1080p or for non-5 second duration @@ -343,7 +347,7 @@ async def api_call( template_id=pixverse_template, seed=seed, ), - auth_kwargs=kwargs, + auth_kwargs=auth, ) response_api = await operation.execute() @@ -364,8 +368,8 @@ async def api_call( PixverseStatus.deleted, ], status_extractor=lambda x: x.Resp.status, - auth_kwargs=kwargs, - node_id=unique_id, + auth_kwargs=auth, + node_id=cls.hidden.unique_id, result_url_extractor=get_video_url_from_response, estimated_duration=AVERAGE_DURATION_I2V, ) @@ -373,72 +377,71 @@ async def api_call( async with aiohttp.ClientSession() as session: async with session.get(response_poll.Resp.url) as vid_response: - return (VideoFromFile(BytesIO(await vid_response.content.read())),) + return comfy_io.NodeOutput(VideoFromFile(BytesIO(await vid_response.content.read()))) -class PixverseTransitionVideoNode(ComfyNodeABC): +class PixverseTransitionVideoNode(comfy_io.ComfyNode): """ Generates videos based on prompt and output_size. """ - RETURN_TYPES = (IO.VIDEO,) - DESCRIPTION = cleandoc(__doc__ or "") # Handle potential None value - FUNCTION = "api_call" - API_NODE = True - CATEGORY = "api node/video/PixVerse" - @classmethod - def INPUT_TYPES(s): - return { - "required": { - "first_frame": (IO.IMAGE,), - "last_frame": (IO.IMAGE,), - "prompt": ( - IO.STRING, - { - "multiline": True, - "default": "", - "tooltip": "Prompt for the video generation", - }, + def define_schema(cls) -> comfy_io.Schema: + return comfy_io.Schema( + node_id="PixverseTransitionVideoNode", + display_name="PixVerse Transition Video", + category="api node/video/PixVerse", + description=cleandoc(cls.__doc__ or ""), + inputs=[ + comfy_io.Image.Input("first_frame"), + comfy_io.Image.Input("last_frame"), + comfy_io.String.Input( + "prompt", + multiline=True, + default="", + tooltip="Prompt for the video generation", ), - "quality": ( - [resolution.value for resolution in PixverseQuality], - { - "default": PixverseQuality.res_540p, - }, + comfy_io.Combo.Input( + "quality", + options=[resolution.value for resolution in PixverseQuality], + default=PixverseQuality.res_540p, ), - "duration_seconds": ([dur.value for dur in PixverseDuration],), - "motion_mode": ([mode.value for mode in PixverseMotionMode],), - "seed": ( - IO.INT, - { - "default": 0, - "min": 0, - "max": 2147483647, - "control_after_generate": True, - "tooltip": "Seed for video generation.", - }, + comfy_io.Combo.Input( + "duration_seconds", + options=[dur.value for dur in PixverseDuration], ), - }, - "optional": { - "negative_prompt": ( - IO.STRING, - { - "default": "", - "forceInput": True, - "tooltip": "An optional text description of undesired elements on an image.", - }, + comfy_io.Combo.Input( + "motion_mode", + options=[mode.value for mode in PixverseMotionMode], ), - }, - "hidden": { - "auth_token": "AUTH_TOKEN_COMFY_ORG", - "comfy_api_key": "API_KEY_COMFY_ORG", - "unique_id": "UNIQUE_ID", - }, - } + comfy_io.Int.Input( + "seed", + default=0, + min=0, + max=2147483647, + control_after_generate=True, + tooltip="Seed for video generation.", + ), + comfy_io.String.Input( + "negative_prompt", + default="", + force_input=True, + tooltip="An optional text description of undesired elements on an image.", + optional=True, + ), + ], + outputs=[comfy_io.Video.Output()], + hidden=[ + comfy_io.Hidden.auth_token_comfy_org, + comfy_io.Hidden.api_key_comfy_org, + comfy_io.Hidden.unique_id, + ], + is_api_node=True, + ) - async def api_call( - self, + @classmethod + async def execute( + cls, first_frame: torch.Tensor, last_frame: torch.Tensor, prompt: str, @@ -447,12 +450,14 @@ async def api_call( motion_mode: str, seed, negative_prompt: str = None, - unique_id: Optional[str] = None, - **kwargs, - ): + ) -> comfy_io.NodeOutput: validate_string(prompt, strip_whitespace=False) - first_frame_id = await upload_image_to_pixverse(first_frame, auth_kwargs=kwargs) - last_frame_id = await upload_image_to_pixverse(last_frame, auth_kwargs=kwargs) + auth = { + "auth_token": cls.hidden.auth_token_comfy_org, + "comfy_api_key": cls.hidden.api_key_comfy_org, + } + first_frame_id = await upload_image_to_pixverse(first_frame, auth_kwargs=auth) + last_frame_id = await upload_image_to_pixverse(last_frame, auth_kwargs=auth) # 1080p is limited to 5 seconds duration # only normal motion_mode supported for 1080p or for non-5 second duration @@ -479,7 +484,7 @@ async def api_call( negative_prompt=negative_prompt if negative_prompt else None, seed=seed, ), - auth_kwargs=kwargs, + auth_kwargs=auth, ) response_api = await operation.execute() @@ -500,8 +505,8 @@ async def api_call( PixverseStatus.deleted, ], status_extractor=lambda x: x.Resp.status, - auth_kwargs=kwargs, - node_id=unique_id, + auth_kwargs=auth, + node_id=cls.hidden.unique_id, result_url_extractor=get_video_url_from_response, estimated_duration=AVERAGE_DURATION_T2V, ) @@ -509,19 +514,19 @@ async def api_call( async with aiohttp.ClientSession() as session: async with session.get(response_poll.Resp.url) as vid_response: - return (VideoFromFile(BytesIO(await vid_response.content.read())),) - - -NODE_CLASS_MAPPINGS = { - "PixverseTextToVideoNode": PixverseTextToVideoNode, - "PixverseImageToVideoNode": PixverseImageToVideoNode, - "PixverseTransitionVideoNode": PixverseTransitionVideoNode, - "PixverseTemplateNode": PixverseTemplateNode, -} - -NODE_DISPLAY_NAME_MAPPINGS = { - "PixverseTextToVideoNode": "PixVerse Text to Video", - "PixverseImageToVideoNode": "PixVerse Image to Video", - "PixverseTransitionVideoNode": "PixVerse Transition Video", - "PixverseTemplateNode": "PixVerse Template", -} + return comfy_io.NodeOutput(VideoFromFile(BytesIO(await vid_response.content.read()))) + + +class PixVerseExtension(ComfyExtension): + @override + async def get_node_list(self) -> list[type[comfy_io.ComfyNode]]: + return [ + PixverseTextToVideoNode, + PixverseImageToVideoNode, + PixverseTransitionVideoNode, + PixverseTemplateNode, + ] + + +async def comfy_entrypoint() -> PixVerseExtension: + return PixVerseExtension() From 5c8e986e273d8af8b976fddbaed726e8278cf1fe Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Fri, 3 Oct 2025 21:50:38 +0300 Subject: [PATCH 25/37] convert nodes_tomesd.py to V3 schema (#10180) --- comfy_extras/nodes_tomesd.py | 50 +++++++++++++++++++++++------------- 1 file changed, 32 insertions(+), 18 deletions(-) diff --git a/comfy_extras/nodes_tomesd.py b/comfy_extras/nodes_tomesd.py index 9f77c06fcb12..87bf29b8fa02 100644 --- a/comfy_extras/nodes_tomesd.py +++ b/comfy_extras/nodes_tomesd.py @@ -1,7 +1,9 @@ #Taken from: https://github.com/dbolya/tomesd import torch -from typing import Tuple, Callable +from typing import Tuple, Callable, Optional +from typing_extensions import override +from comfy_api.latest import ComfyExtension, io import math def do_nothing(x: torch.Tensor, mode:str=None): @@ -144,33 +146,45 @@ def get_functions(x, ratio, original_shape): -class TomePatchModel: +class TomePatchModel(io.ComfyNode): @classmethod - def INPUT_TYPES(s): - return {"required": { "model": ("MODEL",), - "ratio": ("FLOAT", {"default": 0.3, "min": 0.0, "max": 1.0, "step": 0.01}), - }} - RETURN_TYPES = ("MODEL",) - FUNCTION = "patch" + def define_schema(cls): + return io.Schema( + node_id="TomePatchModel", + category="model_patches/unet", + inputs=[ + io.Model.Input("model"), + io.Float.Input("ratio", default=0.3, min=0.0, max=1.0, step=0.01), + ], + outputs=[io.Model.Output()], + ) - CATEGORY = "model_patches/unet" - - def patch(self, model, ratio): - self.u = None + @classmethod + def execute(cls, model, ratio) -> io.NodeOutput: + u: Optional[Callable] = None def tomesd_m(q, k, v, extra_options): + nonlocal u #NOTE: In the reference code get_functions takes x (input of the transformer block) as the argument instead of q #however from my basic testing it seems that using q instead gives better results - m, self.u = get_functions(q, ratio, extra_options["original_shape"]) + m, u = get_functions(q, ratio, extra_options["original_shape"]) return m(q), k, v def tomesd_u(n, extra_options): - return self.u(n) + nonlocal u + return u(n) m = model.clone() m.set_model_attn1_patch(tomesd_m) m.set_model_attn1_output_patch(tomesd_u) - return (m, ) + return io.NodeOutput(m) + + +class TomePatchModelExtension(ComfyExtension): + @override + async def get_node_list(self) -> list[type[io.ComfyNode]]: + return [ + TomePatchModel, + ] -NODE_CLASS_MAPPINGS = { - "TomePatchModel": TomePatchModel, -} +async def comfy_entrypoint() -> TomePatchModelExtension: + return TomePatchModelExtension() From 4614ee09ca1aaca7ee8067d6c5c30695582326ff Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Fri, 3 Oct 2025 23:24:42 +0300 Subject: [PATCH 26/37] convert nodes_edit_model.py to V3 schema (#10147) --- comfy_extras/nodes_edit_model.py | 46 ++++++++++++++++++++------------ 1 file changed, 29 insertions(+), 17 deletions(-) diff --git a/comfy_extras/nodes_edit_model.py b/comfy_extras/nodes_edit_model.py index b69f7971591b..36da66f34e1d 100644 --- a/comfy_extras/nodes_edit_model.py +++ b/comfy_extras/nodes_edit_model.py @@ -1,26 +1,38 @@ import node_helpers +from typing_extensions import override +from comfy_api.latest import ComfyExtension, io -class ReferenceLatent: +class ReferenceLatent(io.ComfyNode): @classmethod - def INPUT_TYPES(s): - return {"required": {"conditioning": ("CONDITIONING", ), - }, - "optional": {"latent": ("LATENT", ),} - } + def define_schema(cls): + return io.Schema( + node_id="ReferenceLatent", + category="advanced/conditioning/edit_models", + description="This node sets the guiding latent for an edit model. If the model supports it you can chain multiple to set multiple reference images.", + inputs=[ + io.Conditioning.Input("conditioning"), + io.Latent.Input("latent", optional=True), + ], + outputs=[ + io.Conditioning.Output(), + ] + ) - RETURN_TYPES = ("CONDITIONING",) - FUNCTION = "append" - - CATEGORY = "advanced/conditioning/edit_models" - DESCRIPTION = "This node sets the guiding latent for an edit model. If the model supports it you can chain multiple to set multiple reference images." - - def append(self, conditioning, latent=None): + @classmethod + def execute(cls, conditioning, latent=None) -> io.NodeOutput: if latent is not None: conditioning = node_helpers.conditioning_set_values(conditioning, {"reference_latents": [latent["samples"]]}, append=True) - return (conditioning, ) + return io.NodeOutput(conditioning) + + +class EditModelExtension(ComfyExtension): + @override + async def get_node_list(self) -> list[type[io.ComfyNode]]: + return [ + ReferenceLatent, + ] -NODE_CLASS_MAPPINGS = { - "ReferenceLatent": ReferenceLatent, -} +def comfy_entrypoint() -> EditModelExtension: + return EditModelExtension() From 93d859cfaaad150c2a1e5e54c8f14765fa79ecb5 Mon Sep 17 00:00:00 2001 From: Finn-Hecker Date: Fri, 3 Oct 2025 23:32:19 +0200 Subject: [PATCH 27/37] Fix type annotation syntax in MotionEncoder_tc __init__ (#10186) ## Summary Fixed incorrect type hint syntax in `MotionEncoder_tc.__init__()` parameter list. ## Changes - Line 647: Changed `num_heads=int` to `num_heads: int` - This corrects the parameter annotation from a default value assignment to proper type hint syntax ## Details The parameter was using assignment syntax (`=`) instead of type annotation syntax (`:`), which would incorrectly set the default value to the `int` class itself rather than annotating the expected type. --- comfy/ldm/wan/model.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/comfy/ldm/wan/model.py b/comfy/ldm/wan/model.py index 0dc650ced357..90c347d3d0b2 100644 --- a/comfy/ldm/wan/model.py +++ b/comfy/ldm/wan/model.py @@ -903,7 +903,7 @@ class MotionEncoder_tc(nn.Module): def __init__(self, in_dim: int, hidden_dim: int, - num_heads=int, + num_heads: int, need_global=True, dtype=None, device=None, From 08726b64fe767f47bf074a05bedd6db45314c4c9 Mon Sep 17 00:00:00 2001 From: comfyanonymous <121283862+comfyanonymous@users.noreply.github.com> Date: Fri, 3 Oct 2025 15:22:43 -0700 Subject: [PATCH 28/37] Update amd nightly command in readme. (#10189) --- README.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index 8f24a33ee2b6..1224a6176a81 100644 --- a/README.md +++ b/README.md @@ -211,9 +211,9 @@ AMD users can install rocm and pytorch with pip if you don't have it already ins ```pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm6.4``` -This is the command to install the nightly with ROCm 6.4 which might have some performance improvements: +This is the command to install the nightly with ROCm 7.0 which might have some performance improvements: -```pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/rocm6.4``` +```pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/rocm7.0``` ### Intel GPUs (Windows and Linux) From bbd683098e7d18700f025b2f0a4f6a44a3176602 Mon Sep 17 00:00:00 2001 From: comfyanonymous <121283862+comfyanonymous@users.noreply.github.com> Date: Fri, 3 Oct 2025 20:37:43 -0700 Subject: [PATCH 29/37] Add instructions to install nightly AMD pytorch for windows. (#10190) * Add instructions to install nightly AMD pytorch for windows. * Update README.md --- README.md | 26 +++++++++++++++++++------- 1 file changed, 19 insertions(+), 7 deletions(-) diff --git a/README.md b/README.md index 1224a6176a81..4a5a17cdae01 100644 --- a/README.md +++ b/README.md @@ -206,7 +206,8 @@ Put your SD checkpoints (the huge ckpt/safetensors files) in: models/checkpoints Put your VAE in: models/vae -### AMD GPUs (Linux only) +### AMD GPUs (Linux) + AMD users can install rocm and pytorch with pip if you don't have it already installed, this is the command to install the stable version: ```pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm6.4``` @@ -215,6 +216,23 @@ This is the command to install the nightly with ROCm 7.0 which might have some p ```pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/rocm7.0``` + +### AMD GPUs (Experimental: Windows and Linux), RDNA 3, 3.5 and 4 only. + +These have less hardware support than the builds above but they work on windows. You also need to install the pytorch version specific to your hardware. + +RDNA 3 (RX 7000 series): + +```pip install --pre torch torchvision torchaudio --index-url https://rocm.nightlies.amd.com/v2/gfx110X-dgpu/``` + +RDNA 3.5 (Strix halo/Ryzen AI Max+ 365): + +```pip install --pre torch torchvision torchaudio --index-url https://rocm.nightlies.amd.com/v2/gfx1151/``` + +RDNA 4 (RX 9000 series): + +```pip install --pre torch torchvision torchaudio --index-url https://rocm.nightlies.amd.com/v2/gfx120X-all/``` + ### Intel GPUs (Windows and Linux) (Option 1) Intel Arc GPU users can install native PyTorch with torch.xpu support using pip. More information can be found [here](https://pytorch.org/docs/main/notes/get_start_xpu.html) @@ -270,12 +288,6 @@ You can install ComfyUI in Apple Mac silicon (M1 or M2) with any recent macOS ve > **Note**: Remember to add your models, VAE, LoRAs etc. to the corresponding Comfy folders, as discussed in [ComfyUI manual installation](#manual-install-windows-linux). -#### DirectML (AMD Cards on Windows) - -This is very badly supported and is not recommended. There are some unofficial builds of pytorch ROCm on windows that exist that will give you a much better experience than this. This readme will be updated once official pytorch ROCm builds for windows come out. - -```pip install torch-directml``` Then you can launch ComfyUI with: ```python main.py --directml``` - #### Ascend NPUs For models compatible with Ascend Extension for PyTorch (torch_npu). To get started, ensure your environment meets the prerequisites outlined on the [installation](https://ascend.github.io/docs/sources/ascend/quick_install.html) page. Here's a step-by-step guide tailored to your platform and installation method: From 22f99fb97edaccf450152c8bf7c4068c1d331899 Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Sat, 4 Oct 2025 22:22:57 +0300 Subject: [PATCH 30/37] fix(api-nodes): enable 2 more pylint rules, removed non needed code (#10192) --- comfy_api_nodes/nodes_gemini.py | 3 +- comfy_api_nodes/nodes_moonvalley.py | 49 ++--------------------------- pyproject.toml | 2 -- 3 files changed, 4 insertions(+), 50 deletions(-) diff --git a/comfy_api_nodes/nodes_gemini.py b/comfy_api_nodes/nodes_gemini.py index 151cb404465b..309e9a2d21e5 100644 --- a/comfy_api_nodes/nodes_gemini.py +++ b/comfy_api_nodes/nodes_gemini.py @@ -39,6 +39,7 @@ tensor_to_base64_string, bytesio_to_image_tensor, ) +from comfy_api.util import VideoContainer, VideoCodec GEMINI_BASE_ENDPOINT = "/proxy/vertexai/gemini" @@ -310,7 +311,7 @@ def create_video_parts(self, video_input: IO.VIDEO, **kwargs) -> list[GeminiPart Returns: List of GeminiPart objects containing the encoded video. """ - from comfy_api.util import VideoContainer, VideoCodec + base_64_string = video_to_base64_string( video_input, container_format=VideoContainer.MP4, diff --git a/comfy_api_nodes/nodes_moonvalley.py b/comfy_api_nodes/nodes_moonvalley.py index 08e838fef65a..6467dd6141a2 100644 --- a/comfy_api_nodes/nodes_moonvalley.py +++ b/comfy_api_nodes/nodes_moonvalley.py @@ -2,11 +2,7 @@ from typing import Any, Callable, Optional, TypeVar import torch from typing_extensions import override -from comfy_api_nodes.util.validation_utils import ( - get_image_dimensions, - validate_image_dimensions, -) - +from comfy_api_nodes.util.validation_utils import validate_image_dimensions from comfy_api_nodes.apis import ( MoonvalleyTextToVideoRequest, @@ -132,47 +128,6 @@ def validate_prompts( return True -def validate_input_media(width, height, with_frame_conditioning, num_frames_in=None): - # inference validation - # T = num_frames - # in all cases, the following must be true: T divisible by 16 and H,W by 8. in addition... - # with image conditioning: H*W must be divisible by 8192 - # without image conditioning: T divisible by 32 - if num_frames_in and not num_frames_in % 16 == 0: - return False, ("The input video total frame count must be divisible by 16!") - - if height % 8 != 0 or width % 8 != 0: - return False, ( - f"Height ({height}) and width ({width}) must be " "divisible by 8" - ) - - if with_frame_conditioning: - if (height * width) % 8192 != 0: - return False, ( - f"Height * width ({height * width}) must be " - "divisible by 8192 for frame conditioning" - ) - else: - if num_frames_in and not num_frames_in % 32 == 0: - return False, ("The input video total frame count must be divisible by 32!") - - -def validate_input_image( - image: torch.Tensor, with_frame_conditioning: bool = False -) -> None: - """ - Validates the input image adheres to the expectations of the API: - - The image resolution should not be less than 300*300px - - The aspect ratio of the image should be between 1:2.5 ~ 2.5:1 - - """ - height, width = get_image_dimensions(image) - validate_input_media(width, height, with_frame_conditioning) - validate_image_dimensions( - image, min_width=300, min_height=300, max_height=MAX_HEIGHT, max_width=MAX_WIDTH - ) - - def validate_video_to_video_input(video: VideoInput) -> VideoInput: """ Validates and processes video input for Moonvalley Video-to-Video generation. @@ -499,7 +454,7 @@ async def execute( seed: int, steps: int, ) -> comfy_io.NodeOutput: - validate_input_image(image, True) + validate_image_dimensions(image, min_width=300, min_height=300, max_height=MAX_HEIGHT, max_width=MAX_WIDTH) validate_prompts(prompt, negative_prompt, MOONVALLEY_MAREY_MAX_PROMPT_LENGTH) width_height = parse_width_height_from_res(resolution) diff --git a/pyproject.toml b/pyproject.toml index 240919a4390e..383e7d10a612 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -70,7 +70,5 @@ messages_control.disable = [ "invalid-overridden-method", "unused-variable", "pointless-string-statement", - "inconsistent-return-statements", - "import-outside-toplevel", "redefined-outer-name", ] From 2ed74f7ac78d3ff713d0a8583695c31055914b76 Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Sat, 4 Oct 2025 22:29:09 +0300 Subject: [PATCH 31/37] convert nodes_rodin.py to V3 schema (#10195) --- comfy_api_nodes/nodes_rodin.py | 915 +++++++++++++++++---------------- 1 file changed, 465 insertions(+), 450 deletions(-) diff --git a/comfy_api_nodes/nodes_rodin.py b/comfy_api_nodes/nodes_rodin.py index 633ac46d335f..bd758f762523 100644 --- a/comfy_api_nodes/nodes_rodin.py +++ b/comfy_api_nodes/nodes_rodin.py @@ -7,14 +7,15 @@ from __future__ import annotations from inspect import cleandoc -from comfy.comfy_types.node_typing import IO import folder_paths as comfy_paths import aiohttp import os import asyncio -import io import logging import math +from typing import Optional +from io import BytesIO +from typing_extensions import override from PIL import Image from comfy_api_nodes.apis.rodin_api import ( Rodin3DGenerateRequest, @@ -31,428 +32,436 @@ SynchronousOperation, PollingOperation, ) +from comfy_api.latest import ComfyExtension, io as comfy_io -COMMON_PARAMETERS = { - "Seed": ( - IO.INT, - { - "default":0, - "min":0, - "max":65535, - "display":"number" - } +COMMON_PARAMETERS = [ + comfy_io.Int.Input( + "Seed", + default=0, + min=0, + max=65535, + display_mode=comfy_io.NumberDisplay.number, + optional=True, ), - "Material_Type": ( - IO.COMBO, - { - "options": ["PBR", "Shaded"], - "default": "PBR" - } + comfy_io.Combo.Input("Material_Type", options=["PBR", "Shaded"], default="PBR", optional=True), + comfy_io.Combo.Input( + "Polygon_count", + options=["4K-Quad", "8K-Quad", "18K-Quad", "50K-Quad", "200K-Triangle"], + default="18K-Quad", + optional=True, ), - "Polygon_count": ( - IO.COMBO, - { - "options": ["4K-Quad", "8K-Quad", "18K-Quad", "50K-Quad", "200K-Triangle"], - "default": "18K-Quad" - } - ) -} +] -def create_task_error(response: Rodin3DGenerateResponse): - """Check if the response has error""" - return hasattr(response, "error") +def get_quality_mode(poly_count): + polycount = poly_count.split("-") + poly = polycount[1] + count = polycount[0] + if poly == "Triangle": + mesh_mode = "Raw" + elif poly == "Quad": + mesh_mode = "Quad" + else: + mesh_mode = "Quad" + + if count == "4K": + quality_override = 4000 + elif count == "8K": + quality_override = 8000 + elif count == "18K": + quality_override = 18000 + elif count == "50K": + quality_override = 50000 + elif count == "2K": + quality_override = 2000 + elif count == "20K": + quality_override = 20000 + elif count == "150K": + quality_override = 150000 + elif count == "500K": + quality_override = 500000 + else: + quality_override = 18000 -class Rodin3DAPI: + return mesh_mode, quality_override + + +def tensor_to_filelike(tensor, max_pixels: int = 2048*2048): """ - Generate 3D Assets using Rodin API + Converts a PyTorch tensor to a file-like object. + + Args: + - tensor (torch.Tensor): A tensor representing an image of shape (H, W, C) + where C is the number of channels (3 for RGB), H is height, and W is width. + + Returns: + - io.BytesIO: A file-like object containing the image data. """ - RETURN_TYPES = (IO.STRING,) - RETURN_NAMES = ("3D Model Path",) - CATEGORY = "api node/3d/Rodin" - DESCRIPTION = cleandoc(__doc__ or "") - FUNCTION = "api_call" - API_NODE = True - - def tensor_to_filelike(self, tensor, max_pixels: int = 2048*2048): - """ - Converts a PyTorch tensor to a file-like object. - - Args: - - tensor (torch.Tensor): A tensor representing an image of shape (H, W, C) - where C is the number of channels (3 for RGB), H is height, and W is width. - - Returns: - - io.BytesIO: A file-like object containing the image data. - """ - array = tensor.cpu().numpy() - array = (array * 255).astype('uint8') - image = Image.fromarray(array, 'RGB') - - original_width, original_height = image.size - original_pixels = original_width * original_height - if original_pixels > max_pixels: - scale = math.sqrt(max_pixels / original_pixels) - new_width = int(original_width * scale) - new_height = int(original_height * scale) - else: - new_width, new_height = original_width, original_height - - if new_width != original_width or new_height != original_height: - image = image.resize((new_width, new_height), Image.Resampling.LANCZOS) - - img_byte_arr = io.BytesIO() - image.save(img_byte_arr, format='PNG') # PNG is used for lossless compression - img_byte_arr.seek(0) - return img_byte_arr - - def check_rodin_status(self, response: Rodin3DCheckStatusResponse) -> str: - has_failed = any(job.status == JobStatus.Failed for job in response.jobs) - all_done = all(job.status == JobStatus.Done for job in response.jobs) - status_list = [str(job.status) for job in response.jobs] - logging.info(f"[ Rodin3D API - CheckStatus ] Generate Status: {status_list}") - if has_failed: - logging.error(f"[ Rodin3D API - CheckStatus ] Generate Failed: {status_list}, Please try again.") - raise Exception("[ Rodin3D API ] Generate Failed, Please Try again.") - elif all_done: - return "DONE" - else: - return "Generating" - - async def create_generate_task(self, images=None, seed=1, material="PBR", quality_override=18000, tier="Regular", mesh_mode="Quad", TAPose = False, **kwargs): - if images is None: - raise Exception("Rodin 3D generate requires at least 1 image.") - if len(images) > 5: - raise Exception("Rodin 3D generate requires up to 5 image.") - - path = "/proxy/rodin/api/v2/rodin" - operation = SynchronousOperation( - endpoint=ApiEndpoint( - path=path, - method=HttpMethod.POST, - request_model=Rodin3DGenerateRequest, - response_model=Rodin3DGenerateResponse, - ), - request=Rodin3DGenerateRequest( - seed=seed, - tier=tier, - material=material, - quality_override=quality_override, - mesh_mode=mesh_mode, - TAPose=TAPose, - ), - files=[ - ( - "images", - open(image, "rb") if isinstance(image, str) else self.tensor_to_filelike(image) - ) - for image in images if image is not None - ], - content_type = "multipart/form-data", - auth_kwargs=kwargs, - ) + array = tensor.cpu().numpy() + array = (array * 255).astype('uint8') + image = Image.fromarray(array, 'RGB') + + original_width, original_height = image.size + original_pixels = original_width * original_height + if original_pixels > max_pixels: + scale = math.sqrt(max_pixels / original_pixels) + new_width = int(original_width * scale) + new_height = int(original_height * scale) + else: + new_width, new_height = original_width, original_height + + if new_width != original_width or new_height != original_height: + image = image.resize((new_width, new_height), Image.Resampling.LANCZOS) + + img_byte_arr = BytesIO() + image.save(img_byte_arr, format='PNG') # PNG is used for lossless compression + img_byte_arr.seek(0) + return img_byte_arr + + +async def create_generate_task( + images=None, + seed=1, + material="PBR", + quality_override=18000, + tier="Regular", + mesh_mode="Quad", + TAPose = False, + auth_kwargs: Optional[dict[str, str]] = None, +): + if images is None: + raise Exception("Rodin 3D generate requires at least 1 image.") + if len(images) > 5: + raise Exception("Rodin 3D generate requires up to 5 image.") + + path = "/proxy/rodin/api/v2/rodin" + operation = SynchronousOperation( + endpoint=ApiEndpoint( + path=path, + method=HttpMethod.POST, + request_model=Rodin3DGenerateRequest, + response_model=Rodin3DGenerateResponse, + ), + request=Rodin3DGenerateRequest( + seed=seed, + tier=tier, + material=material, + quality_override=quality_override, + mesh_mode=mesh_mode, + TAPose=TAPose, + ), + files=[ + ( + "images", + open(image, "rb") if isinstance(image, str) else tensor_to_filelike(image) + ) + for image in images if image is not None + ], + content_type="multipart/form-data", + auth_kwargs=auth_kwargs, + ) - response = await operation.execute() - - if create_task_error(response): - error_message = f"Rodin3D Create 3D generate Task Failed. Message: {response.message}, error: {response.error}" - logging.error(error_message) - raise Exception(error_message) - - logging.info("[ Rodin3D API - Submit Jobs ] Submit Generate Task Success!") - subscription_key = response.jobs.subscription_key - task_uuid = response.uuid - logging.info(f"[ Rodin3D API - Submit Jobs ] UUID: {task_uuid}") - return task_uuid, subscription_key - - async def poll_for_task_status(self, subscription_key, **kwargs) -> Rodin3DCheckStatusResponse: - - path = "/proxy/rodin/api/v2/status" - - poll_operation = PollingOperation( - poll_endpoint=ApiEndpoint( - path = path, - method=HttpMethod.POST, - request_model=Rodin3DCheckStatusRequest, - response_model=Rodin3DCheckStatusResponse, - ), - request=Rodin3DCheckStatusRequest( - subscription_key = subscription_key - ), - completed_statuses=["DONE"], - failed_statuses=["FAILED"], - status_extractor=self.check_rodin_status, - poll_interval=3.0, - auth_kwargs=kwargs, - ) + response = await operation.execute() + + if hasattr(response, "error"): + error_message = f"Rodin3D Create 3D generate Task Failed. Message: {response.message}, error: {response.error}" + logging.error(error_message) + raise Exception(error_message) + + logging.info("[ Rodin3D API - Submit Jobs ] Submit Generate Task Success!") + subscription_key = response.jobs.subscription_key + task_uuid = response.uuid + logging.info(f"[ Rodin3D API - Submit Jobs ] UUID: {task_uuid}") + return task_uuid, subscription_key + + +def check_rodin_status(response: Rodin3DCheckStatusResponse) -> str: + all_done = all(job.status == JobStatus.Done for job in response.jobs) + status_list = [str(job.status) for job in response.jobs] + logging.info(f"[ Rodin3D API - CheckStatus ] Generate Status: {status_list}") + if any(job.status == JobStatus.Failed for job in response.jobs): + logging.error(f"[ Rodin3D API - CheckStatus ] Generate Failed: {status_list}, Please try again.") + raise Exception("[ Rodin3D API ] Generate Failed, Please Try again.") + if all_done: + return "DONE" + return "Generating" + + +async def poll_for_task_status( + subscription_key, auth_kwargs: Optional[dict[str, str]] = None, +) -> Rodin3DCheckStatusResponse: + poll_operation = PollingOperation( + poll_endpoint=ApiEndpoint( + path="/proxy/rodin/api/v2/status", + method=HttpMethod.POST, + request_model=Rodin3DCheckStatusRequest, + response_model=Rodin3DCheckStatusResponse, + ), + request=Rodin3DCheckStatusRequest(subscription_key=subscription_key), + completed_statuses=["DONE"], + failed_statuses=["FAILED"], + status_extractor=check_rodin_status, + poll_interval=3.0, + auth_kwargs=auth_kwargs, + ) + logging.info("[ Rodin3D API - CheckStatus ] Generate Start!") + return await poll_operation.execute() + + +async def get_rodin_download_list(uuid, auth_kwargs: Optional[dict[str, str]] = None) -> Rodin3DDownloadResponse: + logging.info("[ Rodin3D API - Downloading ] Generate Successfully!") + operation = SynchronousOperation( + endpoint=ApiEndpoint( + path="/proxy/rodin/api/v2/download", + method=HttpMethod.POST, + request_model=Rodin3DDownloadRequest, + response_model=Rodin3DDownloadResponse, + ), + request=Rodin3DDownloadRequest(task_uuid=uuid), + auth_kwargs=auth_kwargs, + ) + return await operation.execute() + + +async def download_files(url_list, task_uuid): + save_path = os.path.join(comfy_paths.get_output_directory(), f"Rodin3D_{task_uuid}") + os.makedirs(save_path, exist_ok=True) + model_file_path = None + async with aiohttp.ClientSession() as session: + for i in url_list.list: + url = i.url + file_name = i.name + file_path = os.path.join(save_path, file_name) + if file_path.endswith(".glb"): + model_file_path = file_path + logging.info(f"[ Rodin3D API - download_files ] Downloading file: {file_path}") + max_retries = 5 + for attempt in range(max_retries): + try: + async with session.get(url) as resp: + resp.raise_for_status() + with open(file_path, "wb") as f: + async for chunk in resp.content.iter_chunked(32 * 1024): + f.write(chunk) + break + except Exception as e: + logging.info(f"[ Rodin3D API - download_files ] Error downloading {file_path}:{e}") + if attempt < max_retries - 1: + logging.info("Retrying...") + await asyncio.sleep(2) + else: + logging.info( + "[ Rodin3D API - download_files ] Failed to download %s after %s attempts.", + file_path, + max_retries, + ) + return model_file_path + + +class Rodin3D_Regular(comfy_io.ComfyNode): + """Generate 3D Assets using Rodin API""" - logging.info("[ Rodin3D API - CheckStatus ] Generate Start!") - - return await poll_operation.execute() - - async def get_rodin_download_list(self, uuid, **kwargs) -> Rodin3DDownloadResponse: - logging.info("[ Rodin3D API - Downloading ] Generate Successfully!") - - path = "/proxy/rodin/api/v2/download" - operation = SynchronousOperation( - endpoint=ApiEndpoint( - path=path, - method=HttpMethod.POST, - request_model=Rodin3DDownloadRequest, - response_model=Rodin3DDownloadResponse, - ), - request=Rodin3DDownloadRequest( - task_uuid=uuid - ), - auth_kwargs=kwargs + @classmethod + def define_schema(cls) -> comfy_io.Schema: + return comfy_io.Schema( + node_id="Rodin3D_Regular", + display_name="Rodin 3D Generate - Regular Generate", + category="api node/3d/Rodin", + description=cleandoc(cls.__doc__ or ""), + inputs=[ + comfy_io.Image.Input("Images"), + *COMMON_PARAMETERS, + ], + outputs=[comfy_io.String.Output(display_name="3D Model Path")], + hidden=[ + comfy_io.Hidden.auth_token_comfy_org, + comfy_io.Hidden.api_key_comfy_org, + ], + is_api_node=True, ) - return await operation.execute() - - def get_quality_mode(self, poly_count): - polycount = poly_count.split("-") - poly = polycount[1] - count = polycount[0] - if poly == "Triangle": - mesh_mode = "Raw" - elif poly == "Quad": - mesh_mode = "Quad" - else: - mesh_mode = "Quad" - - if count == "4K": - quality_override = 4000 - elif count == "8K": - quality_override = 8000 - elif count == "18K": - quality_override = 18000 - elif count == "50K": - quality_override = 50000 - elif count == "2K": - quality_override = 2000 - elif count == "20K": - quality_override = 20000 - elif count == "150K": - quality_override = 150000 - elif count == "500K": - quality_override = 500000 - else: - quality_override = 18000 - - return mesh_mode, quality_override - - async def download_files(self, url_list, task_uuid): - save_path = os.path.join(comfy_paths.get_output_directory(), f"Rodin3D_{task_uuid}") - os.makedirs(save_path, exist_ok=True) - model_file_path = None - async with aiohttp.ClientSession() as session: - for i in url_list.list: - url = i.url - file_name = i.name - file_path = os.path.join(save_path, file_name) - if file_path.endswith(".glb"): - model_file_path = file_path - logging.info(f"[ Rodin3D API - download_files ] Downloading file: {file_path}") - max_retries = 5 - for attempt in range(max_retries): - try: - async with session.get(url) as resp: - resp.raise_for_status() - with open(file_path, "wb") as f: - async for chunk in resp.content.iter_chunked(32 * 1024): - f.write(chunk) - break - except Exception as e: - logging.info(f"[ Rodin3D API - download_files ] Error downloading {file_path}:{e}") - if attempt < max_retries - 1: - logging.info("Retrying...") - await asyncio.sleep(2) - else: - logging.info( - "[ Rodin3D API - download_files ] Failed to download %s after %s attempts.", - file_path, - max_retries, - ) - - return model_file_path - - -class Rodin3D_Regular(Rodin3DAPI): @classmethod - def INPUT_TYPES(s): - return { - "required": { - "Images": - ( - IO.IMAGE, - { - "forceInput":True, - } - ) - }, - "optional": { - **COMMON_PARAMETERS - }, - "hidden": { - "auth_token": "AUTH_TOKEN_COMFY_ORG", - "comfy_api_key": "API_KEY_COMFY_ORG", - }, - } - - async def api_call( - self, + async def execute( + cls, Images, Seed, Material_Type, Polygon_count, - **kwargs - ): + ) -> comfy_io.NodeOutput: tier = "Regular" num_images = Images.shape[0] m_images = [] for i in range(num_images): m_images.append(Images[i]) - mesh_mode, quality_override = self.get_quality_mode(Polygon_count) - task_uuid, subscription_key = await self.create_generate_task(images=m_images, seed=Seed, material=Material_Type, - quality_override=quality_override, tier=tier, mesh_mode=mesh_mode, - **kwargs) - await self.poll_for_task_status(subscription_key, **kwargs) - download_list = await self.get_rodin_download_list(task_uuid, **kwargs) - model = await self.download_files(download_list, task_uuid) + mesh_mode, quality_override = get_quality_mode(Polygon_count) + auth = { + "auth_token": cls.hidden.auth_token_comfy_org, + "comfy_api_key": cls.hidden.api_key_comfy_org, + } + task_uuid, subscription_key = await create_generate_task( + images=m_images, + seed=Seed, + material=Material_Type, + quality_override=quality_override, + tier=tier, + mesh_mode=mesh_mode, + auth_kwargs=auth, + ) + await poll_for_task_status(subscription_key, auth_kwargs=auth) + download_list = await get_rodin_download_list(task_uuid, auth_kwargs=auth) + model = await download_files(download_list, task_uuid) + + return comfy_io.NodeOutput(model) - return (model,) +class Rodin3D_Detail(comfy_io.ComfyNode): + """Generate 3D Assets using Rodin API""" -class Rodin3D_Detail(Rodin3DAPI): @classmethod - def INPUT_TYPES(s): - return { - "required": { - "Images": - ( - IO.IMAGE, - { - "forceInput":True, - } - ) - }, - "optional": { - **COMMON_PARAMETERS - }, - "hidden": { - "auth_token": "AUTH_TOKEN_COMFY_ORG", - "comfy_api_key": "API_KEY_COMFY_ORG", - }, - } + def define_schema(cls) -> comfy_io.Schema: + return comfy_io.Schema( + node_id="Rodin3D_Detail", + display_name="Rodin 3D Generate - Detail Generate", + category="api node/3d/Rodin", + description=cleandoc(cls.__doc__ or ""), + inputs=[ + comfy_io.Image.Input("Images"), + *COMMON_PARAMETERS, + ], + outputs=[comfy_io.String.Output(display_name="3D Model Path")], + hidden=[ + comfy_io.Hidden.auth_token_comfy_org, + comfy_io.Hidden.api_key_comfy_org, + ], + is_api_node=True, + ) - async def api_call( - self, + @classmethod + async def execute( + cls, Images, Seed, Material_Type, Polygon_count, - **kwargs - ): + ) -> comfy_io.NodeOutput: tier = "Detail" num_images = Images.shape[0] m_images = [] for i in range(num_images): m_images.append(Images[i]) - mesh_mode, quality_override = self.get_quality_mode(Polygon_count) - task_uuid, subscription_key = await self.create_generate_task(images=m_images, seed=Seed, material=Material_Type, - quality_override=quality_override, tier=tier, mesh_mode=mesh_mode, - **kwargs) - await self.poll_for_task_status(subscription_key, **kwargs) - download_list = await self.get_rodin_download_list(task_uuid, **kwargs) - model = await self.download_files(download_list, task_uuid) + mesh_mode, quality_override = get_quality_mode(Polygon_count) + auth = { + "auth_token": cls.hidden.auth_token_comfy_org, + "comfy_api_key": cls.hidden.api_key_comfy_org, + } + task_uuid, subscription_key = await create_generate_task( + images=m_images, + seed=Seed, + material=Material_Type, + quality_override=quality_override, + tier=tier, + mesh_mode=mesh_mode, + auth_kwargs=auth, + ) + await poll_for_task_status(subscription_key, auth_kwargs=auth) + download_list = await get_rodin_download_list(task_uuid, auth_kwargs=auth) + model = await download_files(download_list, task_uuid) + + return comfy_io.NodeOutput(model) - return (model,) +class Rodin3D_Smooth(comfy_io.ComfyNode): + """Generate 3D Assets using Rodin API""" -class Rodin3D_Smooth(Rodin3DAPI): @classmethod - def INPUT_TYPES(s): - return { - "required": { - "Images": - ( - IO.IMAGE, - { - "forceInput":True, - } - ) - }, - "optional": { - **COMMON_PARAMETERS - }, - "hidden": { - "auth_token": "AUTH_TOKEN_COMFY_ORG", - "comfy_api_key": "API_KEY_COMFY_ORG", - }, - } + def define_schema(cls) -> comfy_io.Schema: + return comfy_io.Schema( + node_id="Rodin3D_Smooth", + display_name="Rodin 3D Generate - Smooth Generate", + category="api node/3d/Rodin", + description=cleandoc(cls.__doc__ or ""), + inputs=[ + comfy_io.Image.Input("Images"), + *COMMON_PARAMETERS, + ], + outputs=[comfy_io.String.Output(display_name="3D Model Path")], + hidden=[ + comfy_io.Hidden.auth_token_comfy_org, + comfy_io.Hidden.api_key_comfy_org, + ], + is_api_node=True, + ) - async def api_call( - self, + @classmethod + async def execute( + cls, Images, Seed, Material_Type, Polygon_count, - **kwargs - ): + ) -> comfy_io.NodeOutput: tier = "Smooth" num_images = Images.shape[0] m_images = [] for i in range(num_images): m_images.append(Images[i]) - mesh_mode, quality_override = self.get_quality_mode(Polygon_count) - task_uuid, subscription_key = await self.create_generate_task(images=m_images, seed=Seed, material=Material_Type, - quality_override=quality_override, tier=tier, mesh_mode=mesh_mode, - **kwargs) - await self.poll_for_task_status(subscription_key, **kwargs) - download_list = await self.get_rodin_download_list(task_uuid, **kwargs) - model = await self.download_files(download_list, task_uuid) + mesh_mode, quality_override = get_quality_mode(Polygon_count) + auth = { + "auth_token": cls.hidden.auth_token_comfy_org, + "comfy_api_key": cls.hidden.api_key_comfy_org, + } + task_uuid, subscription_key = await create_generate_task( + images=m_images, + seed=Seed, + material=Material_Type, + quality_override=quality_override, + tier=tier, + mesh_mode=mesh_mode, + auth_kwargs=auth, + ) + await poll_for_task_status(subscription_key, auth_kwargs=auth) + download_list = await get_rodin_download_list(task_uuid, auth_kwargs=auth) + model = await download_files(download_list, task_uuid) + + return comfy_io.NodeOutput(model) - return (model,) +class Rodin3D_Sketch(comfy_io.ComfyNode): + """Generate 3D Assets using Rodin API""" -class Rodin3D_Sketch(Rodin3DAPI): @classmethod - def INPUT_TYPES(s): - return { - "required": { - "Images": - ( - IO.IMAGE, - { - "forceInput":True, - } - ) - }, - "optional": { - "Seed": - ( - IO.INT, - { - "default":0, - "min":0, - "max":65535, - "display":"number" - } - ) - }, - "hidden": { - "auth_token": "AUTH_TOKEN_COMFY_ORG", - "comfy_api_key": "API_KEY_COMFY_ORG", - }, - } + def define_schema(cls) -> comfy_io.Schema: + return comfy_io.Schema( + node_id="Rodin3D_Sketch", + display_name="Rodin 3D Generate - Sketch Generate", + category="api node/3d/Rodin", + description=cleandoc(cls.__doc__ or ""), + inputs=[ + comfy_io.Image.Input("Images"), + comfy_io.Int.Input( + "Seed", + default=0, + min=0, + max=65535, + display_mode=comfy_io.NumberDisplay.number, + optional=True, + ), + ], + outputs=[comfy_io.String.Output(display_name="3D Model Path")], + hidden=[ + comfy_io.Hidden.auth_token_comfy_org, + comfy_io.Hidden.api_key_comfy_org, + ], + is_api_node=True, + ) - async def api_call( - self, + @classmethod + async def execute( + cls, Images, Seed, - **kwargs - ): + ) -> comfy_io.NodeOutput: tier = "Sketch" num_images = Images.shape[0] m_images = [] @@ -461,104 +470,110 @@ async def api_call( material_type = "PBR" quality_override = 18000 mesh_mode = "Quad" - task_uuid, subscription_key = await self.create_generate_task( - images=m_images, seed=Seed, material=material_type, quality_override=quality_override, tier=tier, mesh_mode=mesh_mode, **kwargs + auth = { + "auth_token": cls.hidden.auth_token_comfy_org, + "comfy_api_key": cls.hidden.api_key_comfy_org, + } + task_uuid, subscription_key = await create_generate_task( + images=m_images, + seed=Seed, + material=material_type, + quality_override=quality_override, + tier=tier, + mesh_mode=mesh_mode, + auth_kwargs=auth, ) - await self.poll_for_task_status(subscription_key, **kwargs) - download_list = await self.get_rodin_download_list(task_uuid, **kwargs) - model = await self.download_files(download_list, task_uuid) + await poll_for_task_status(subscription_key, auth_kwargs=auth) + download_list = await get_rodin_download_list(task_uuid, auth_kwargs=auth) + model = await download_files(download_list, task_uuid) + + return comfy_io.NodeOutput(model) - return (model,) -class Rodin3D_Gen2(Rodin3DAPI): +class Rodin3D_Gen2(comfy_io.ComfyNode): + """Generate 3D Assets using Rodin API""" + @classmethod - def INPUT_TYPES(s): - return { - "required": { - "Images": - ( - IO.IMAGE, - { - "forceInput":True, - } - ) - }, - "optional": { - "Seed": ( - IO.INT, - { - "default":0, - "min":0, - "max":65535, - "display":"number" - } - ), - "Material_Type": ( - IO.COMBO, - { - "options": ["PBR", "Shaded"], - "default": "PBR" - } + def define_schema(cls) -> comfy_io.Schema: + return comfy_io.Schema( + node_id="Rodin3D_Gen2", + display_name="Rodin 3D Generate - Gen-2 Generate", + category="api node/3d/Rodin", + description=cleandoc(cls.__doc__ or ""), + inputs=[ + comfy_io.Image.Input("Images"), + comfy_io.Int.Input( + "Seed", + default=0, + min=0, + max=65535, + display_mode=comfy_io.NumberDisplay.number, + optional=True, ), - "Polygon_count": ( - IO.COMBO, - { - "options": ["4K-Quad", "8K-Quad", "18K-Quad", "50K-Quad", "2K-Triangle", "20K-Triangle", "150K-Triangle", "500K-Triangle"], - "default": "500K-Triangle" - } + comfy_io.Combo.Input("Material_Type", options=["PBR", "Shaded"], default="PBR", optional=True), + comfy_io.Combo.Input( + "Polygon_count", + options=["4K-Quad", "8K-Quad", "18K-Quad", "50K-Quad", "2K-Triangle", "20K-Triangle", "150K-Triangle", "500K-Triangle"], + default="500K-Triangle", + optional=True, ), - "TAPose": ( - IO.BOOLEAN, - { - "default": False, - } - ) - }, - "hidden": { - "auth_token": "AUTH_TOKEN_COMFY_ORG", - "comfy_api_key": "API_KEY_COMFY_ORG", - }, - } + comfy_io.Boolean.Input("TAPose", default=False), + ], + outputs=[comfy_io.String.Output(display_name="3D Model Path")], + hidden=[ + comfy_io.Hidden.auth_token_comfy_org, + comfy_io.Hidden.api_key_comfy_org, + ], + is_api_node=True, + ) - async def api_call( - self, + @classmethod + async def execute( + cls, Images, Seed, Material_Type, Polygon_count, TAPose, - **kwargs - ): + ) -> comfy_io.NodeOutput: tier = "Gen-2" num_images = Images.shape[0] m_images = [] for i in range(num_images): m_images.append(Images[i]) - mesh_mode, quality_override = self.get_quality_mode(Polygon_count) - task_uuid, subscription_key = await self.create_generate_task(images=m_images, seed=Seed, material=Material_Type, - quality_override=quality_override, tier=tier, mesh_mode=mesh_mode, TAPose=TAPose, - **kwargs) - await self.poll_for_task_status(subscription_key, **kwargs) - download_list = await self.get_rodin_download_list(task_uuid, **kwargs) - model = await self.download_files(download_list, task_uuid) - - return (model,) - -# A dictionary that contains all nodes you want to export with their names -# NOTE: names should be globally unique -NODE_CLASS_MAPPINGS = { - "Rodin3D_Regular": Rodin3D_Regular, - "Rodin3D_Detail": Rodin3D_Detail, - "Rodin3D_Smooth": Rodin3D_Smooth, - "Rodin3D_Sketch": Rodin3D_Sketch, - "Rodin3D_Gen2": Rodin3D_Gen2, -} - -# A dictionary that contains the friendly/humanly readable titles for the nodes -NODE_DISPLAY_NAME_MAPPINGS = { - "Rodin3D_Regular": "Rodin 3D Generate - Regular Generate", - "Rodin3D_Detail": "Rodin 3D Generate - Detail Generate", - "Rodin3D_Smooth": "Rodin 3D Generate - Smooth Generate", - "Rodin3D_Sketch": "Rodin 3D Generate - Sketch Generate", - "Rodin3D_Gen2": "Rodin 3D Generate - Gen-2 Generate", -} + mesh_mode, quality_override = get_quality_mode(Polygon_count) + auth = { + "auth_token": cls.hidden.auth_token_comfy_org, + "comfy_api_key": cls.hidden.api_key_comfy_org, + } + task_uuid, subscription_key = await create_generate_task( + images=m_images, + seed=Seed, + material=Material_Type, + quality_override=quality_override, + tier=tier, + mesh_mode=mesh_mode, + TAPose=TAPose, + auth_kwargs=auth, + ) + await poll_for_task_status(subscription_key, auth_kwargs=auth) + download_list = await get_rodin_download_list(task_uuid, auth_kwargs=auth) + model = await download_files(download_list, task_uuid) + + return comfy_io.NodeOutput(model) + + +class Rodin3DExtension(ComfyExtension): + @override + async def get_node_list(self) -> list[type[comfy_io.ComfyNode]]: + return [ + Rodin3D_Regular, + Rodin3D_Detail, + Rodin3D_Smooth, + Rodin3D_Sketch, + Rodin3D_Gen2, + ] + + +async def comfy_entrypoint() -> Rodin3DExtension: + return Rodin3DExtension() From b1fa1922df597af759150f4e26ecb276c9753ee4 Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Sat, 4 Oct 2025 22:33:48 +0300 Subject: [PATCH 32/37] convert nodes_stable3d.py to V3 schema (#10204) --- comfy_extras/nodes_stable3d.py | 155 +++++++++++++++++++-------------- 1 file changed, 89 insertions(+), 66 deletions(-) diff --git a/comfy_extras/nodes_stable3d.py b/comfy_extras/nodes_stable3d.py index be2e34c28f49..c6d8a683d5f9 100644 --- a/comfy_extras/nodes_stable3d.py +++ b/comfy_extras/nodes_stable3d.py @@ -1,6 +1,8 @@ import torch import nodes import comfy.utils +from typing_extensions import override +from comfy_api.latest import ComfyExtension, io def camera_embeddings(elevation, azimuth): elevation = torch.as_tensor([elevation]) @@ -20,26 +22,31 @@ def camera_embeddings(elevation, azimuth): return embeddings -class StableZero123_Conditioning: +class StableZero123_Conditioning(io.ComfyNode): @classmethod - def INPUT_TYPES(s): - return {"required": { "clip_vision": ("CLIP_VISION",), - "init_image": ("IMAGE",), - "vae": ("VAE",), - "width": ("INT", {"default": 256, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 8}), - "height": ("INT", {"default": 256, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 8}), - "batch_size": ("INT", {"default": 1, "min": 1, "max": 4096}), - "elevation": ("FLOAT", {"default": 0.0, "min": -180.0, "max": 180.0, "step": 0.1, "round": False}), - "azimuth": ("FLOAT", {"default": 0.0, "min": -180.0, "max": 180.0, "step": 0.1, "round": False}), - }} - RETURN_TYPES = ("CONDITIONING", "CONDITIONING", "LATENT") - RETURN_NAMES = ("positive", "negative", "latent") - - FUNCTION = "encode" - - CATEGORY = "conditioning/3d_models" - - def encode(self, clip_vision, init_image, vae, width, height, batch_size, elevation, azimuth): + def define_schema(cls): + return io.Schema( + node_id="StableZero123_Conditioning", + category="conditioning/3d_models", + inputs=[ + io.ClipVision.Input("clip_vision"), + io.Image.Input("init_image"), + io.Vae.Input("vae"), + io.Int.Input("width", default=256, min=16, max=nodes.MAX_RESOLUTION, step=8), + io.Int.Input("height", default=256, min=16, max=nodes.MAX_RESOLUTION, step=8), + io.Int.Input("batch_size", default=1, min=1, max=4096), + io.Float.Input("elevation", default=0.0, min=-180.0, max=180.0, step=0.1, round=False), + io.Float.Input("azimuth", default=0.0, min=-180.0, max=180.0, step=0.1, round=False) + ], + outputs=[ + io.Conditioning.Output(display_name="positive"), + io.Conditioning.Output(display_name="negative"), + io.Latent.Output(display_name="latent") + ] + ) + + @classmethod + def execute(cls, clip_vision, init_image, vae, width, height, batch_size, elevation, azimuth) -> io.NodeOutput: output = clip_vision.encode_image(init_image) pooled = output.image_embeds.unsqueeze(0) pixels = comfy.utils.common_upscale(init_image.movedim(-1,1), width, height, "bilinear", "center").movedim(1,-1) @@ -51,30 +58,35 @@ def encode(self, clip_vision, init_image, vae, width, height, batch_size, elevat positive = [[cond, {"concat_latent_image": t}]] negative = [[torch.zeros_like(pooled), {"concat_latent_image": torch.zeros_like(t)}]] latent = torch.zeros([batch_size, 4, height // 8, width // 8]) - return (positive, negative, {"samples":latent}) + return io.NodeOutput(positive, negative, {"samples":latent}) + +class StableZero123_Conditioning_Batched(io.ComfyNode): + @classmethod + def define_schema(cls): + return io.Schema( + node_id="StableZero123_Conditioning_Batched", + category="conditioning/3d_models", + inputs=[ + io.ClipVision.Input("clip_vision"), + io.Image.Input("init_image"), + io.Vae.Input("vae"), + io.Int.Input("width", default=256, min=16, max=nodes.MAX_RESOLUTION, step=8), + io.Int.Input("height", default=256, min=16, max=nodes.MAX_RESOLUTION, step=8), + io.Int.Input("batch_size", default=1, min=1, max=4096), + io.Float.Input("elevation", default=0.0, min=-180.0, max=180.0, step=0.1, round=False), + io.Float.Input("azimuth", default=0.0, min=-180.0, max=180.0, step=0.1, round=False), + io.Float.Input("elevation_batch_increment", default=0.0, min=-180.0, max=180.0, step=0.1, round=False), + io.Float.Input("azimuth_batch_increment", default=0.0, min=-180.0, max=180.0, step=0.1, round=False) + ], + outputs=[ + io.Conditioning.Output(display_name="positive"), + io.Conditioning.Output(display_name="negative"), + io.Latent.Output(display_name="latent") + ] + ) -class StableZero123_Conditioning_Batched: @classmethod - def INPUT_TYPES(s): - return {"required": { "clip_vision": ("CLIP_VISION",), - "init_image": ("IMAGE",), - "vae": ("VAE",), - "width": ("INT", {"default": 256, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 8}), - "height": ("INT", {"default": 256, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 8}), - "batch_size": ("INT", {"default": 1, "min": 1, "max": 4096}), - "elevation": ("FLOAT", {"default": 0.0, "min": -180.0, "max": 180.0, "step": 0.1, "round": False}), - "azimuth": ("FLOAT", {"default": 0.0, "min": -180.0, "max": 180.0, "step": 0.1, "round": False}), - "elevation_batch_increment": ("FLOAT", {"default": 0.0, "min": -180.0, "max": 180.0, "step": 0.1, "round": False}), - "azimuth_batch_increment": ("FLOAT", {"default": 0.0, "min": -180.0, "max": 180.0, "step": 0.1, "round": False}), - }} - RETURN_TYPES = ("CONDITIONING", "CONDITIONING", "LATENT") - RETURN_NAMES = ("positive", "negative", "latent") - - FUNCTION = "encode" - - CATEGORY = "conditioning/3d_models" - - def encode(self, clip_vision, init_image, vae, width, height, batch_size, elevation, azimuth, elevation_batch_increment, azimuth_batch_increment): + def execute(cls, clip_vision, init_image, vae, width, height, batch_size, elevation, azimuth, elevation_batch_increment, azimuth_batch_increment) -> io.NodeOutput: output = clip_vision.encode_image(init_image) pooled = output.image_embeds.unsqueeze(0) pixels = comfy.utils.common_upscale(init_image.movedim(-1,1), width, height, "bilinear", "center").movedim(1,-1) @@ -93,27 +105,32 @@ def encode(self, clip_vision, init_image, vae, width, height, batch_size, elevat positive = [[cond, {"concat_latent_image": t}]] negative = [[torch.zeros_like(pooled), {"concat_latent_image": torch.zeros_like(t)}]] latent = torch.zeros([batch_size, 4, height // 8, width // 8]) - return (positive, negative, {"samples":latent, "batch_index": [0] * batch_size}) + return io.NodeOutput(positive, negative, {"samples":latent, "batch_index": [0] * batch_size}) -class SV3D_Conditioning: +class SV3D_Conditioning(io.ComfyNode): @classmethod - def INPUT_TYPES(s): - return {"required": { "clip_vision": ("CLIP_VISION",), - "init_image": ("IMAGE",), - "vae": ("VAE",), - "width": ("INT", {"default": 576, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 8}), - "height": ("INT", {"default": 576, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 8}), - "video_frames": ("INT", {"default": 21, "min": 1, "max": 4096}), - "elevation": ("FLOAT", {"default": 0.0, "min": -90.0, "max": 90.0, "step": 0.1, "round": False}), - }} - RETURN_TYPES = ("CONDITIONING", "CONDITIONING", "LATENT") - RETURN_NAMES = ("positive", "negative", "latent") - - FUNCTION = "encode" - - CATEGORY = "conditioning/3d_models" - - def encode(self, clip_vision, init_image, vae, width, height, video_frames, elevation): + def define_schema(cls): + return io.Schema( + node_id="SV3D_Conditioning", + category="conditioning/3d_models", + inputs=[ + io.ClipVision.Input("clip_vision"), + io.Image.Input("init_image"), + io.Vae.Input("vae"), + io.Int.Input("width", default=576, min=16, max=nodes.MAX_RESOLUTION, step=8), + io.Int.Input("height", default=576, min=16, max=nodes.MAX_RESOLUTION, step=8), + io.Int.Input("video_frames", default=21, min=1, max=4096), + io.Float.Input("elevation", default=0.0, min=-90.0, max=90.0, step=0.1, round=False) + ], + outputs=[ + io.Conditioning.Output(display_name="positive"), + io.Conditioning.Output(display_name="negative"), + io.Latent.Output(display_name="latent") + ] + ) + + @classmethod + def execute(cls, clip_vision, init_image, vae, width, height, video_frames, elevation) -> io.NodeOutput: output = clip_vision.encode_image(init_image) pooled = output.image_embeds.unsqueeze(0) pixels = comfy.utils.common_upscale(init_image.movedim(-1,1), width, height, "bilinear", "center").movedim(1,-1) @@ -133,11 +150,17 @@ def encode(self, clip_vision, init_image, vae, width, height, video_frames, elev positive = [[pooled, {"concat_latent_image": t, "elevation": elevations, "azimuth": azimuths}]] negative = [[torch.zeros_like(pooled), {"concat_latent_image": torch.zeros_like(t), "elevation": elevations, "azimuth": azimuths}]] latent = torch.zeros([video_frames, 4, height // 8, width // 8]) - return (positive, negative, {"samples":latent}) + return io.NodeOutput(positive, negative, {"samples":latent}) + +class Stable3DExtension(ComfyExtension): + @override + async def get_node_list(self) -> list[type[io.ComfyNode]]: + return [ + StableZero123_Conditioning, + StableZero123_Conditioning_Batched, + SV3D_Conditioning, + ] -NODE_CLASS_MAPPINGS = { - "StableZero123_Conditioning": StableZero123_Conditioning, - "StableZero123_Conditioning_Batched": StableZero123_Conditioning_Batched, - "SV3D_Conditioning": SV3D_Conditioning, -} +async def comfy_entrypoint() -> Stable3DExtension: + return Stable3DExtension() From caf07331ff1b20f4104b9693ed244d6e22f80b5a Mon Sep 17 00:00:00 2001 From: comfyanonymous <121283862+comfyanonymous@users.noreply.github.com> Date: Sat, 4 Oct 2025 19:05:05 -0700 Subject: [PATCH 33/37] Remove soundfile dependency. No more torchaudio load or save. (#10210) --- comfy_extras/nodes_audio.py | 2 +- requirements.txt | 1 - 2 files changed, 1 insertion(+), 2 deletions(-) diff --git a/comfy_extras/nodes_audio.py b/comfy_extras/nodes_audio.py index 51c8b9dd93cd..1c868fcbab7e 100644 --- a/comfy_extras/nodes_audio.py +++ b/comfy_extras/nodes_audio.py @@ -360,7 +360,7 @@ def INPUT_TYPES(s): def load(self, audio): audio_path = folder_paths.get_annotated_filepath(audio) - waveform, sample_rate = torchaudio.load(audio_path) + waveform, sample_rate = load(audio_path) audio = {"waveform": waveform.unsqueeze(0), "sample_rate": sample_rate} return (audio, ) diff --git a/requirements.txt b/requirements.txt index 588c5dcf01cc..6c28f9478c8e 100644 --- a/requirements.txt +++ b/requirements.txt @@ -25,6 +25,5 @@ av>=14.2.0 #non essential dependencies: kornia>=0.7.1 spandrel -soundfile pydantic~=2.0 pydantic-settings~=2.0 From 187f43696dd58f252075d2e3c6873706eb6b5fa1 Mon Sep 17 00:00:00 2001 From: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Date: Sun, 5 Oct 2025 09:34:18 +0300 Subject: [PATCH 34/37] fix(api-nodes): disable "std" mode for Kling2.5-turbo (#10212) --- comfy_api_nodes/nodes_kling.py | 3 +++ 1 file changed, 3 insertions(+) diff --git a/comfy_api_nodes/nodes_kling.py b/comfy_api_nodes/nodes_kling.py index d8646f1060ca..44fccc0c7498 100644 --- a/comfy_api_nodes/nodes_kling.py +++ b/comfy_api_nodes/nodes_kling.py @@ -712,6 +712,9 @@ async def api_call( # Camera control type for image 2 video is always `simple` camera_control.type = KlingCameraControlType.simple + if mode == "std" and model_name == KlingVideoGenModelName.kling_v2_5_turbo.value: + mode = "pro" # October 5: currently "std" mode is not supported for this model + initial_operation = SynchronousOperation( endpoint=ApiEndpoint( path=PATH_IMAGE_TO_VIDEO, From 195e0b063950f585fe584c5ce7b0b689f8d20ff8 Mon Sep 17 00:00:00 2001 From: comfyanonymous <121283862+comfyanonymous@users.noreply.github.com> Date: Sun, 5 Oct 2025 12:41:19 -0700 Subject: [PATCH 35/37] Remove useless code. (#10223) --- comfy/ldm/ace/vae/music_dcae_pipeline.py | 11 ----------- 1 file changed, 11 deletions(-) diff --git a/comfy/ldm/ace/vae/music_dcae_pipeline.py b/comfy/ldm/ace/vae/music_dcae_pipeline.py index af81280eb0dd..3c8830c176f7 100644 --- a/comfy/ldm/ace/vae/music_dcae_pipeline.py +++ b/comfy/ldm/ace/vae/music_dcae_pipeline.py @@ -23,8 +23,6 @@ def __init__(self, source_sample_rate=None, dcae_config={}, vocoder_config={}): else: self.source_sample_rate = source_sample_rate - # self.resampler = torchaudio.transforms.Resample(source_sample_rate, 44100) - self.transform = transforms.Compose([ transforms.Normalize(0.5, 0.5), ]) @@ -37,10 +35,6 @@ def __init__(self, source_sample_rate=None, dcae_config={}, vocoder_config={}): self.scale_factor = 0.1786 self.shift_factor = -1.9091 - def load_audio(self, audio_path): - audio, sr = torchaudio.load(audio_path) - return audio, sr - def forward_mel(self, audios): mels = [] for i in range(len(audios)): @@ -73,10 +67,8 @@ def encode(self, audios, audio_lengths=None, sr=None): latent = self.dcae.encoder(mel.unsqueeze(0)) latents.append(latent) latents = torch.cat(latents, dim=0) - # latent_lengths = (audio_lengths / sr * 44100 / 512 / self.time_dimention_multiple).long() latents = (latents - self.shift_factor) * self.scale_factor return latents - # return latents, latent_lengths @torch.no_grad() def decode(self, latents, audio_lengths=None, sr=None): @@ -91,9 +83,7 @@ def decode(self, latents, audio_lengths=None, sr=None): wav = self.vocoder.decode(mels[0]).squeeze(1) if sr is not None: - # resampler = torchaudio.transforms.Resample(44100, sr).to(latents.device).to(latents.dtype) wav = torchaudio.functional.resample(wav, 44100, sr) - # wav = resampler(wav) else: sr = 44100 pred_wavs.append(wav) @@ -101,7 +91,6 @@ def decode(self, latents, audio_lengths=None, sr=None): if audio_lengths is not None: pred_wavs = [wav[:, :length].cpu() for wav, length in zip(pred_wavs, audio_lengths)] return torch.stack(pred_wavs) - # return sr, pred_wavs def forward(self, audios, audio_lengths=None, sr=None): latents, latent_lengths = self.encode(audios=audios, audio_lengths=audio_lengths, sr=sr) From 7326e46deeab97219cad32d0624991f9ffea4fe5 Mon Sep 17 00:00:00 2001 From: ComfyUI Wiki Date: Tue, 7 Oct 2025 01:57:00 +0800 Subject: [PATCH 36/37] Update template to 0.1.93 (#10235) * Update template to 0.1.92 * Update template to 0.1.93 --- requirements.txt | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/requirements.txt b/requirements.txt index 6c28f9478c8e..db0486960217 100644 --- a/requirements.txt +++ b/requirements.txt @@ -1,5 +1,5 @@ comfyui-frontend-package==1.27.7 -comfyui-workflow-templates==0.1.91 +comfyui-workflow-templates==0.1.93 comfyui-embedded-docs==0.2.6 torch torchsde From 6bd3f8eb9ff2d7c74e8ca75ad1f854a6b256b714 Mon Sep 17 00:00:00 2001 From: comfyanonymous Date: Mon, 6 Oct 2025 14:49:04 -0400 Subject: [PATCH 37/37] ComfyUI version 0.3.63 --- comfyui_version.py | 2 +- pyproject.toml | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/comfyui_version.py b/comfyui_version.py index ac76fbe358f5..c3257d4bff58 100644 --- a/comfyui_version.py +++ b/comfyui_version.py @@ -1,3 +1,3 @@ # This file is automatically generated by the build process when version is # updated in pyproject.toml. -__version__ = "0.3.62" +__version__ = "0.3.63" diff --git a/pyproject.toml b/pyproject.toml index 383e7d10a612..a9e3de0c6140 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -1,6 +1,6 @@ [project] name = "ComfyUI" -version = "0.3.62" +version = "0.3.63" readme = "README.md" license = { file = "LICENSE" } requires-python = ">=3.9"