🚀 扁平色彩风格模型
本项目的扁平色彩风格模型专注于生成无明显线条、色彩扁平且深度感较弱的图像和视频,为图像与视频生成领域带来独特的视觉体验。
🚀 快速开始
触发词使用
- 使用
flat color
触发图像生成。
- 使用
no lineart
触发图像生成。
模型下载
本模型的权重以 Safetensors 格式提供,可在 Files & versions 标签页 中下载。
✨ 主要特性
- 独特风格:基于无明显线条、扁平色彩和弱深度感的图像进行训练,生成具有独特风格的图像和视频。
- 广泛应用:适用于多种场景,如虚拟主播形象、动漫风格视频等。
📦 安装指南
加载 LoRA 时,使用 LoraLoaderModelOnly
节点,并使用 fp16
的 wan2.1_t2v_1.3B_fp16.safetensors
。
💻 使用示例
基础用法
可在 ComfyUI_examples/wan/#text-to-video 中查看文本到视频的预览示例。
以下是一些示例输入及对应输出:
示例 1
输入文本:flat color, no lineart, blending, negative space, artist:[john kafka|ponsuke kaikai|hara id 21|yoneyama mai|fuzichoco], 1girl, hoshimachi suisei, virtual youtuber, blue hair, side ponytail, cowboy shot, black shirt, star print, off shoulder, outdoors, starry sky, wariza, looking up, half-closed eyes, black sky, live2d animation, upper body, high quality cinematic video of a woman sitting under the starry night sky. The Camera is steady, This is a cowboy shot. The animation is smooth and fluid.
负提示词:bad quality video,色调艳丽,过曝,静态,细节模糊不清,字幕,风格,作品,画作,画面,静止,整体发灰,最差质量,低质量,JPEG压缩残留,丑陋的,残缺的,多余的手指,画得不好的手部,画得不好的脸部,畸形的,毁容的,形态畸形的肢体,手指融合,静止不动的画面,杂乱的背景,三条腿,背景人很多,倒着走
输出:[images/ComfyUI_00455_.webp](images/ComfyUI_00455_.webp)
示例 2
输入文本:flat color, no lineart, blending, negative space, artist:[john kafka|ponsuke kaikai|hara id 21|yoneyama mai|fuzichoco], 1girl, sakura miko, pink hair, cowboy shot, white shirt, floral print, off shoulder, outdoors, cherry blossom, tree shade, wariza, looking up, falling petals, half-closed eyes, white sky, clouds, live2d animation, upper body, high quality cinematic video of a woman sitting under a sakura tree. Dreamy and lonely, the camera close-ups on the face of the woman as she turns towards the viewer. The Camera is steady, This is a cowboy shot. The animation is smooth and fluid.
负提示词:bad quality video,色调艳丽,过曝,静态,细节模糊不清,字幕,风格,作品,画作,画面,静止,整体发灰,最差质量,低质量,JPEG压缩残留,丑陋的,残缺的,多余的手指,画得不好的手部,画得不好的脸部,畸形的,毁容的,形态畸形的肢体,手指融合,静止不动的画面,杂乱的背景,三条腿,背景人很多,倒着走
输出:[images/ComfyUI_00469_.webp](images/ComfyUI_00469_.webp)
📚 详细文档
模型描述
本模型基于 Wan-AI/Wan2.1-T2V-1.3B-Diffusers 基础模型进行训练。相关内容转载自 CivitAI。
训练配置
本模型使用 diffusion-pipe 进行训练,以下是详细的训练配置文件:
dataset.toml
resolutions = [512]
enable_ar_bucket = true
min_ar = 0.5
max_ar = 2.0
num_ar_buckets = 7
frame_buckets = [1]
[[directory]]
path = '/mnt/d/huanvideo/training_data/images'
num_repeats = 5
resolutions = [720]
frame_buckets = [1]
[[directory]]
path = '/mnt/d/huanvideo/training_data/videos'
num_repeats = 5
resolutions = [512]
frame_buckets = [6, 28, 31, 32, 36, 42, 43, 48, 50, 53]
config.toml
output_dir = '/mnt/d/wan/training_output'
dataset = 'dataset.toml'
epochs = 50
micro_batch_size_per_gpu = 1
pipeline_stages = 1
gradient_accumulation_steps = 4
gradient_clipping = 1.0
warmup_steps = 100
eval_every_n_epochs = 5
eval_before_first_step = true
eval_micro_batch_size_per_gpu = 1
eval_gradient_accumulation_steps = 1
save_every_n_epochs = 5
checkpoint_every_n_minutes = 30
activation_checkpointing = true
partition_method = 'parameters'
save_dtype = 'bfloat16'
caching_batch_size = 1
steps_per_print = 1
video_clip_mode = 'single_middle'
[model]
type = 'wan'
ckpt_path = '../Wan2.1-T2V-1.3B'
dtype = 'bfloat16'
transformer_dtype = 'float8'
timestep_sample_method = 'logit_normal'
[adapter]
type = 'lora'
rank = 32
dtype = 'bfloat16'
[optimizer]
type = 'adamw_optimi'
lr = 5e-5
betas = [0.9, 0.99]
weight_decay = 0.02
eps = 1e-8
📄 许可证
本项目采用 Apache-2.0 许可证。