From 9cf299a9f9488e4cb9b3f7cef3bc94c185c19f73 Mon Sep 17 00:00:00 2001 From: comfyanonymous <121283862+comfyanonymous@users.noreply.github.com> Date: Fri, 23 Jan 2026 16:50:48 -0800 Subject: [PATCH] Make regular empty latent node work properly on flux 2 variants. (#12050) --- comfy/latent_formats.py | 3 +++ comfy/sample.py | 12 +++++++++--- comfy_extras/nodes_custom_sampler.py | 6 ++++-- comfy_extras/nodes_sd3.py | 2 +- nodes.py | 5 +++-- 5 files changed, 20 insertions(+), 8 deletions(-) diff --git a/comfy/latent_formats.py b/comfy/latent_formats.py index cb4f52ce1..5600825ed 100644 --- a/comfy/latent_formats.py +++ b/comfy/latent_formats.py @@ -8,6 +8,7 @@ class LatentFormat: latent_rgb_factors_bias = None latent_rgb_factors_reshape = None taesd_decoder_name = None + spacial_downscale_ratio = 8 def process_in(self, latent): return latent * self.scale_factor @@ -181,6 +182,7 @@ class Flux(SD3): class Flux2(LatentFormat): latent_channels = 128 + spacial_downscale_ratio = 16 def __init__(self): self.latent_rgb_factors =[ @@ -749,6 +751,7 @@ class ACEAudio(LatentFormat): class ChromaRadiance(LatentFormat): latent_channels = 3 + spacial_downscale_ratio = 1 def __init__(self): self.latent_rgb_factors = [ diff --git a/comfy/sample.py b/comfy/sample.py index 2f8f3a51c..a2a39b527 100644 --- a/comfy/sample.py +++ b/comfy/sample.py @@ -37,12 +37,18 @@ def prepare_noise(latent_image, seed, noise_inds=None): return noises -def fix_empty_latent_channels(model, latent_image): +def fix_empty_latent_channels(model, latent_image, downscale_ratio_spacial=None): if latent_image.is_nested: return latent_image latent_format = model.get_model_object("latent_format") #Resize the empty latent image so it has the right number of channels - if latent_format.latent_channels != latent_image.shape[1] and torch.count_nonzero(latent_image) == 0: - latent_image = comfy.utils.repeat_to_batch_size(latent_image, latent_format.latent_channels, dim=1) + if torch.count_nonzero(latent_image) == 0: + if latent_format.latent_channels != latent_image.shape[1]: + latent_image = comfy.utils.repeat_to_batch_size(latent_image, latent_format.latent_channels, dim=1) + if downscale_ratio_spacial is not None: + if downscale_ratio_spacial != latent_format.spacial_downscale_ratio: + ratio = downscale_ratio_spacial / latent_format.spacial_downscale_ratio + latent_image = comfy.utils.common_upscale(latent_image, round(latent_image.shape[-1] * ratio), round(latent_image.shape[-2] * ratio), "nearest-exact", crop="disabled") + if latent_format.latent_dimensions == 3 and latent_image.ndim == 4: latent_image = latent_image.unsqueeze(2) return latent_image diff --git a/comfy_extras/nodes_custom_sampler.py b/comfy_extras/nodes_custom_sampler.py index 3eb40e937..a4d84ddf7 100644 --- a/comfy_extras/nodes_custom_sampler.py +++ b/comfy_extras/nodes_custom_sampler.py @@ -741,7 +741,7 @@ class SamplerCustom(io.ComfyNode): latent = latent_image latent_image = latent["samples"] latent = latent.copy() - latent_image = comfy.sample.fix_empty_latent_channels(model, latent_image) + latent_image = comfy.sample.fix_empty_latent_channels(model, latent_image, latent.get("downscale_ratio_spacial", None)) latent["samples"] = latent_image if not add_noise: @@ -760,6 +760,7 @@ class SamplerCustom(io.ComfyNode): samples = comfy.sample.sample_custom(model, noise, cfg, sampler, sigmas, positive, negative, latent_image, noise_mask=noise_mask, callback=callback, disable_pbar=disable_pbar, seed=noise_seed) out = latent.copy() + out.pop("downscale_ratio_spacial", None) out["samples"] = samples if "x0" in x0_output: x0_out = model.model.process_latent_out(x0_output["x0"].cpu()) @@ -939,7 +940,7 @@ class SamplerCustomAdvanced(io.ComfyNode): latent = latent_image latent_image = latent["samples"] latent = latent.copy() - latent_image = comfy.sample.fix_empty_latent_channels(guider.model_patcher, latent_image) + latent_image = comfy.sample.fix_empty_latent_channels(guider.model_patcher, latent_image, latent.get("downscale_ratio_spacial", None)) latent["samples"] = latent_image noise_mask = None @@ -954,6 +955,7 @@ class SamplerCustomAdvanced(io.ComfyNode): samples = samples.to(comfy.model_management.intermediate_device()) out = latent.copy() + out.pop("downscale_ratio_spacial", None) out["samples"] = samples if "x0" in x0_output: x0_out = guider.model_patcher.model.process_latent_out(x0_output["x0"].cpu()) diff --git a/comfy_extras/nodes_sd3.py b/comfy_extras/nodes_sd3.py index 02e5e7dd8..736213a47 100644 --- a/comfy_extras/nodes_sd3.py +++ b/comfy_extras/nodes_sd3.py @@ -55,7 +55,7 @@ class EmptySD3LatentImage(io.ComfyNode): @classmethod def execute(cls, width, height, batch_size=1) -> io.NodeOutput: latent = torch.zeros([batch_size, 16, height // 8, width // 8], device=comfy.model_management.intermediate_device()) - return io.NodeOutput({"samples":latent}) + return io.NodeOutput({"samples": latent, "downscale_ratio_spacial": 8}) generate = execute # TODO: remove diff --git a/nodes.py b/nodes.py index 158106686..b75247665 100644 --- a/nodes.py +++ b/nodes.py @@ -1230,7 +1230,7 @@ class EmptyLatentImage: def generate(self, width, height, batch_size=1): latent = torch.zeros([batch_size, 4, height // 8, width // 8], device=self.device) - return ({"samples":latent}, ) + return ({"samples": latent, "downscale_ratio_spacial": 8}, ) class LatentFromBatch: @@ -1538,7 +1538,7 @@ class SetLatentNoiseMask: def common_ksampler(model, seed, steps, cfg, sampler_name, scheduler, positive, negative, latent, denoise=1.0, disable_noise=False, start_step=None, last_step=None, force_full_denoise=False): latent_image = latent["samples"] - latent_image = comfy.sample.fix_empty_latent_channels(model, latent_image) + latent_image = comfy.sample.fix_empty_latent_channels(model, latent_image, latent.get("downscale_ratio_spacial", None)) if disable_noise: noise = torch.zeros(latent_image.size(), dtype=latent_image.dtype, layout=latent_image.layout, device="cpu") @@ -1556,6 +1556,7 @@ def common_ksampler(model, seed, steps, cfg, sampler_name, scheduler, positive, denoise=denoise, disable_noise=disable_noise, start_step=start_step, last_step=last_step, force_full_denoise=force_full_denoise, noise_mask=noise_mask, callback=callback, disable_pbar=disable_pbar, seed=seed) out = latent.copy() + out.pop("downscale_ratio_spacial", None) out["samples"] = samples return (out, )