🚀 Flux-Isometric-Site-LoRA模型
Flux-Isometric-Site-LoRA是一个用于图像生成的模型，通过特定的参数设置和触发词，能够生成等轴测建筑风格的图像，为图像创作提供了新的可能。
🚀 快速开始
环境设置
import torch
from pipelines import DiffusionPipeline
base_model = "black-forest-labs/FLUX.1-dev"
pipe = DiffusionPipeline.from_pretrained(base_model, torch_dtype=torch.bfloat16)
lora_repo = "strangerzonehf/Flux-Isometric-Site-LoRA"
trigger_word = "Isometric Building"
pipe.load_lora_weights(lora_repo)
device = torch.device("cuda")
pipe.to(device)
触发词使用
你应该使用 Isometric Building 来触发图像生成。
模型下载
该模型的权重以Safetensors格式提供。
点击下载，可在 Files & versions 标签中获取。
✨ 主要特性
- 支持文本到图像的转换，通过特定的触发词生成等轴测建筑风格的图像。
- 提供了丰富的图像生成示例，展示了不同场景下的等轴测建筑图像。
📦 安装指南
本部分代码展示了如何设置环境以使用该模型：
import torch
from pipelines import DiffusionPipeline
base_model = "black-forest-labs/FLUX.1-dev"
pipe = DiffusionPipeline.from_pretrained(base_model, torch_dtype=torch.bfloat16)
lora_repo = "strangerzonehf/Flux-Isometric-Site-LoRA"
trigger_word = "Isometric Building"
pipe.load_lora_weights(lora_repo)
device = torch.device("cuda")
pipe.to(device)
💻 使用示例
基础用法
import torch
from pipelines import DiffusionPipeline
base_model = "black-forest-labs/FLUX.1-dev"
pipe = DiffusionPipeline.from_pretrained(base_model, torch_dtype=torch.bfloat16)
lora_repo = "strangerzonehf/Flux-Isometric-Site-LoRA"
trigger_word = "Isometric Building"
pipe.load_lora_weights(lora_repo)
device = torch.device("cuda")
pipe.to(device)
📚 详细文档
模型描述
图像生成示例
| 输入文本 |
输出图像 |
| 'Isometric Building: A top-down view of a rustic farmhouse with a green tin roof and red-brick walls. The farm includes a small barn, a chicken coop, and a tractor parked in the field. Rows of crops stretch out behind the farmhouse, and a wooden fence surrounds the property.' |
点击查看 |
| 'Isometric Building: A top-down view of a two-story beach house with a large wooden deck overlooking the ocean. The house features a blue roof and white walls, with large glass doors opening to the deck. The deck has lounge chairs, a surfboard, and a barbecue grill. The sandy beach stretches out in the background, with palm trees swaying in the breeze.' |
点击查看 |
| 'Isometric Building, a 3-level view of a two-story house, adorned with a black roof and a black balcony. The house is situated on a grassy lawn, with a swimming pool on the right side of the frame. The pool is surrounded by a wooden deck, with chairs and a white umbrella on the deck. A black SUV is parked in front of the house, adding a pop of color to the scene. A fire pit is lit up in the middle of the lawn, creating a warm and inviting ambiance.' |
点击查看 |
| 'Isometric Building, a top-down view of a two-story white house with a blue roof is seen against a beige backdrop. The house is adorned with a variety of plants, including a tree, a table, and a bench, adding a touch of greenery to the scene. The staircase leads to the entrance of the house, which is flanked by a brown wooden gate. The ground is covered in grass, with a small patio in the middle of the stairs.' |
点击查看 |
| 'Isometric Building, a top-down view of a two-story house with a pool in the foreground. The pool is surrounded by a lush green lawn, with a variety of plants and shrubs, adding a touch of greenery to the scene. The house is adorned with a white bed, a white blanket, and a white umbrella. The balcony of the house casts a shadow on the ground, providing a contrast to the blue sky.' |
点击查看 |
| 'Isometric Building, An aerial view of a two story white house with a blue pool in the middle of it. There is a white car parked on the right side of the house. There are two trees on the left side and right side as well as a small white car on the bottom right side.' |
点击查看 |
图像生成参数
| 参数 |
值 |
参数 |
值 |
| LR Scheduler |
constant |
Noise Offset |
0.03 |
| Optimizer |
AdamW |
Multires Noise Discount |
0.1 |
| Network Dim |
64 |
Multires Noise Iterations |
10 |
| Network Alpha |
32 |
Repeat & Steps |
24 & 2770 |
| Epoch |
17 |
Save Every N Epochs |
1 |
标注信息
标注使用：florence2 - en（自然语言 & 英语）
训练数据
总共使用了 18 张高清图像进行训练。
最佳尺寸与推理
| 尺寸 |
长宽比 |
推荐情况 |
| 1280 x 832 |
3:2 |
最佳 |
| 1024 x 1024 |
1:1 |
默认 |
推理范围
🔧 技术细节
本模型基于 black-forest-labs/FLUX.1-dev 基础模型，使用 LoRA 技术进行微调。在训练过程中，使用了特定的图像生成参数，如 LR Scheduler 设置为 constant，Optimizer 选择 AdamW 等。通过这些参数的设置，使得模型能够生成高质量的等轴测建筑图像。同时，模型在不同的尺寸和长宽比下有不同的表现，推荐的尺寸和长宽比能够获得更好的生成效果。
📄 许可证
本模型使用 creativeml-openrail-m 许可证。
%20--%3e%3cdefs%3e%3cstyle%3e%20.st0%20{%20fill:%20%23061b40;%20}%20.st1%20{%20fill:%20%23306af1;%20}%20.st2%20{%20fill:%20%235ce5cf;%20}%20%3c/style%3e%3c/defs%3e%3cg%3e%3cpath%20class='st0'%20d='M55,10.5h9v3h-9v9h12V7.5h-12v3ZM64,19.5h-6v-3h6v3Z'/%3e%3cpolygon%20class='st0'%20points='69%2016.5%2078%2016.5%2078%2019.5%2069%2019.5%2069%2022.5%2081%2022.5%2081%2013.5%2072%2013.5%2072%2010.5%2081%2010.5%2081%207.5%2069%207.5%2069%2016.5'/%3e%3cpolygon%20class='st0'%20points='95%2010.5%2095%207.5%2083%207.5%2083%2022.5%2095%2022.5%2095%2019.5%2086%2019.5%2086%2016.5%2095%2016.5%2095%2013.5%2086%2013.5%2086%2010.5%2095%2010.5'/%3e%3cpath%20class='st0'%20d='M40,1.5v21h11.6l1.4-1.4v-7.6h0c0,0-1.4-1.5-1.4-1.5l1.4-1.4V2.9l-1.4-1.4h-11.6ZM50,19.5h-7v-6h7v6ZM50,10.5h-7v-6h7v6Z'/%3e%3c/g%3e%3cpath%20class='st1'%20d='M23.1,24L14.7,4.8l-4.9,11.2h3.8l-1.8,4H3.3L12.1,0H2C.9,0,0,.9,0,2v20c0,1.1.9,2,2,2h21.1Z'/%3e%3cpath%20class='st2'%20d='M34,0h-16.8l10.6,24h6.2c1.1,0,2-.9,2-2V2C36,.9,35.1,0,34,0ZM32.5,20h-4V4h4v16Z'/%3e%3c/svg%3e)
Transformers 支持多种语言