🚀 電影感照片生成模型 - cinematicShot - Pics - Flux
本模型專注於生成具有電影感的照片,通過特定的參數設置和觸發詞,能夠根據文本描述生成高質量的圖像,為圖像創作提供了新的可能。
🚀 快速開始
模型設置
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/cinematicShot-Pics-Flux"
trigger_word = "cinematic shot"
pipe.load_lora_weights(lora_repo)
device = torch.device("cuda")
pipe.to(device)
觸發圖像生成
你可以使用 cinematic shot
作為觸發詞來生成圖像。
模型下載
本模型的權重以 Safetensors 格式提供。你可以在 Files & versions 標籤頁 下載。
✨ 主要特性
- 基於
black - forest - labs/FLUX.1 - dev
基礎模型,結合 LoRA 技術進行微調。
- 支持使用
cinematic shot
觸發詞生成具有電影感的圖像。
- 提供多種圖像生成示例,涵蓋不同場景和人物形象。
📦 安裝指南
安裝所需的庫:
import torch
from pipelines import DiffusionPipeline
💻 使用示例
基礎用法
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/cinematicShot-Pics-Flux"
trigger_word = "cinematic shot"
pipe.load_lora_weights(lora_repo)
device = torch.device("cuda")
pipe.to(device)
prompt = 'cinematic shot, Captured at eye - level on a cloudy day, a woman stands with her arms outstretched, wearing a brown turtleneck and blue jeans. Her jacket is draped over her shoulders, adding a touch of warmth to her outfit. The jacket is adorned with a white fur trim. The womans hair is pulled back in a ponytail, framing her face. In the distance, a body of water can be seen, with a mountain range in the background.'
image = pipe(prompt).images[0]
image.save("output_image.png")
📚 詳細文檔
圖像生成參數
屬性 |
詳情 |
基礎模型 |
black - forest - labs/FLUX.1 - dev |
觸發提示詞 |
cinematic shot |
許可證 |
creativeml - openrail - m |
圖像生成示例
文本提示 |
輸出圖像 |
cinematic shot, Captured at eye - level on a cloudy day, a woman stands with her arms outstretched, wearing a brown turtleneck and blue jeans. Her jacket is draped over her shoulders, adding a touch of warmth to her outfit. The jacket is adorned with a white fur trim. The womans hair is pulled back in a ponytail, framing her face. In the distance, a body of water can be seen, with a mountain range in the background. |
[點擊查看](https://cdn - uploads.huggingface.co/production/uploads/65bb837dbfb878f46c77de4c/sfTr7x9n4dUxHAj952aVN.png) |
cinematic shot, a medium - angle shot captures a young woman seated in a convertible car. She is dressed in a red and blue sweater, a white scarf, and blue jeans. Her left arm is raised above her head, adding a touch of balance to the scene. The cars interior is a deep blue, with a black seat, and a white canopy. The womans gaze is directed towards the left side of the frame, and her left hand is resting on her right arm. |
點擊查看 |
cinematic shot, Captured at eye - level on a high - angle shot of a young man sitting on the hood of a white sports car. The man is dressed in a black and white jacket, black pants, and white sneakers. His hair is short and dark, and he has a serious expression on his face. The car is adorned with the word Color in bold black letters, and the back of the car is a bright orange. The sky behind the man is a light blue with a few wispy clouds. |
點擊查看 |
cinematic shot, a medium - sized man stands in the foreground of the frame. He is dressed in a black jacket, black pants, and a black backpack. His hair is short and dark, and hes looking off to the side. The backdrop is a stark white wall, and the ceiling is adorned with metal pipes and beams. To the left of the man, a man with dark hair and glasses is looking off into the distance. |
點擊查看 |
最佳尺寸與推理
尺寸 |
縱橫比 |
推薦情況 |
1280 x 832 |
3:2 |
最佳 |
1024 x 1024 |
1:1 |
默認 |
推理範圍
🔧 技術細節
圖像處理參數
參數 |
值 |
參數 |
值 |
LR 調度器 |
constant |
噪聲偏移 |
0.03 |
優化器 |
AdamW |
多分辨率噪聲折扣 |
0.1 |
網絡維度 |
64 |
多分辨率噪聲迭代次數 |
10 |
網絡 Alpha |
32 |
重複與步數 |
17 & 2600 |
訓練輪數 |
22 |
每 N 輪保存一次 |
1 |
標註信息
標註使用 florence2 - en
(自然語言 & 英語)。
訓練圖像數量
總共使用 33 張圖像進行訓練。
📄 許可證
本模型使用 creativeml - openrail - m
許可證。