98 lines
3.4 KiB
Python
98 lines
3.4 KiB
Python
|
|
from nodes import VAEDecode, VAEEncode
|
|
import comfy
|
|
from .utils import image_2dtransform, apply_easing, easing_functions
|
|
import torch
|
|
from tqdm import tqdm
|
|
|
|
|
|
|
|
|
|
|
|
class Easy2DDeforum:
|
|
|
|
@classmethod
|
|
def INPUT_TYPES(s):
|
|
return {
|
|
"required": {
|
|
"model": ("MODEL", ),
|
|
"vae": ("VAE", ),
|
|
"positive": ("CONDITIONING", ),
|
|
"negative": ("CONDITIONING", ),
|
|
"image": ("IMAGE", ),
|
|
"seed": ("INT", {"default": 0, "min": 0, "max": 0xffffffffffffffff}),
|
|
"steps": ("INT", {"default": 20, "min": 1, "max": 10000}),
|
|
"cfg": ("FLOAT", {"default": 8.0, "min": 0.0, "max": 100.0, "step":0.1, "round": 0.01}),
|
|
"sampler_name": (comfy.samplers.KSampler.SAMPLERS, ),
|
|
"scheduler": (comfy.samplers.KSampler.SCHEDULERS, ),
|
|
"frame": ("INT", {"default": 16}),
|
|
"x": ("INT", {"default": 15, "step": 1, "min": -4096, "max": 4096}),
|
|
"y": ("INT", {"default": 0, "step": 1, "min": -4096, "max": 4096}),
|
|
"zoom": ("FLOAT", {"default": 0.98, "min": 0.001, "step": 0.01}),
|
|
"angle": ("INT", {"default": -1, "step": 1, "min": -360, "max": 360}),
|
|
"denoise_min": ("FLOAT", {"default": 0.40, "min": 0.00, "max": 1.00, "step":0.01}),
|
|
"denoise_max": ("FLOAT", {"default": 0.60, "min": 0.00, "max": 1.00, "step":0.01}),
|
|
"easing_type": (list(easing_functions.keys()), ),
|
|
}
|
|
}
|
|
|
|
RETURN_TYPES = ("IMAGE", )
|
|
RETURN_NAMES = ("IMAGE", )
|
|
FUNCTION = "apply"
|
|
|
|
CATEGORY = "easyDeforum"
|
|
|
|
def apply(self, model, vae, positive, negative, image, frame, seed, steps, cfg, sampler_name, scheduler, x, y, zoom, angle, denoise_min, denoise_max, easing_type):
|
|
|
|
# 初始化模型
|
|
vaedecode = VAEDecode()
|
|
vaeencode = VAEEncode()
|
|
|
|
res = [image]
|
|
|
|
pbar = comfy.utils.ProgressBar(frame)
|
|
for i in tqdm(range(frame)):
|
|
|
|
# 计算噪声
|
|
denoise = (denoise_max - denoise_min) * apply_easing((i+1)/frame, easing_type) + denoise_min
|
|
|
|
# 变换
|
|
image = image_2dtransform(image, x, y, zoom, angle, 0, "reflect")
|
|
|
|
# 编码
|
|
latent = vaeencode.encode(vae, image)[0]
|
|
|
|
|
|
# 计算噪声
|
|
noise = comfy.sample.prepare_noise(latent["samples"], i, None)
|
|
|
|
# 采样
|
|
samples = comfy.sample.sample(model, noise, steps, cfg, sampler_name, scheduler, positive, negative, latent["samples"],
|
|
denoise=denoise, disable_noise=False, start_step=None, last_step=None,
|
|
force_full_denoise=False, noise_mask=None, callback=None, disable_pbar=True, seed=seed+i)
|
|
|
|
# 解码
|
|
image = vaedecode.decode(vae, {"samples": samples})[0]
|
|
|
|
# 这里还能加预览图片
|
|
pbar.update_absolute(i + 1, frame, None)
|
|
|
|
res.append(image)
|
|
|
|
# 如果第一张图片和生成的图片尺寸不一致，则丢弃
|
|
if res[0].size() != res[-1].size():
|
|
res = res[1:]
|
|
|
|
res = torch.cat(res, dim=0)
|
|
return (res, )
|
|
|
|
|
|
# NODE MAPPING
|
|
NODE_CLASS_MAPPINGS = {
|
|
"Easy2DDeforum": Easy2DDeforum,
|
|
}
|
|
|
|
NODE_DISPLAY_NAME_MAPPINGS = {
|
|
"Easy2DDeforum": "Easy2DDeforum (Chan)"
|
|
}
|