My stuff

  • My Workflows

  • Liked Workflows

  • Following Workflows

Go to OpenArt main site
Upload workflow

In Context lora + Character story generator + flux+ shichen

5.0

0 reviews
18
4.6K
1.2K
1
Description
  • # Random Character Movie Story Generator
  • ## Project Overview
  • This is a cloud-based AI movie story generation system that automatically creates movie posters and key stills based on user-input characters and scenes. The entire system adopts a modular design, primarily leveraging StepFun and BizyAir platforms.
  • ## Technical Architecture
  • ### Core Components
  • - **Base Platforms**:
  •  - StepFun Cloud Platform (https://platform.stepfun.com/account-overview)
  •  - BizyAir Workflow Engine (https://github.com/siliconflow/bizyair)
  • ### Image Generation Pipeline
  • 1. **Initial Image Generation**:
  •   - Uses Pulid model for initial character image generation
  •  
  • 2. **Continuity Assurance**:
  •   - Implements In-Context-LoRA technology (https://github.com/ali-vilab/In-Context-LoRA)
  •   - Ensures consistency in generated image style
  •   - Produces key story frame sequences
  • 3. **Final Output**:
  •   - Generates movie poster
  •   - Outputs key scene stills
  • ## Deployment Options
  • - **Cloud Deployment**:
  •  - Default usage of StepFun platform cloud nodes
  •  - No local configuration required
  •  
  • - **Local Deployment**:
  •  - Nodes can be replaced with local ones based on hardware specifications
  •  - Supports hybrid deployment mode
  • ## Workflow
  • 1. User uploads character information
  • 2. Inputs scene description and story outline
  • 3. System automatically generates initial character image
  • 4. Creates continuous scenes through LoRA technology
  • 5. Final synthesis of movie poster
  • ## Key Features
  • - Full cloud operation, zero local dependencies
  • - Modular design supporting custom extensions
  • - Flux-based stable workflow
  • - Highly customizable deployment solutions
  • # 随机角色电影故事生成器
  • ## 项目概述
  • 这是一个基于云平台的AI电影故事生成系统,可以根据用户输入的角色和场景,自动生成电影海报和关键剧照。整个系统采用模块化设计,主要依托于阶跃星辰(StepFun)和BizyAir平台。
  • ## 技术架构
  • ### 核心组件
  • - **基础平台**:
  •  - StepFun云平台 (https://platform.stepfun.com/account-overview)
  •  - BizyAir工作流引擎 (https://github.com/siliconflow/bizyair)
  • ### 图像生成流程
  • 1. **首图生成**:
  •   - 使用Pulid模型生成初始角色图像
  •  
  • 2. **连续性保证**:
  •   - 采用In-Context-LoRA技术 (https://github.com/ali-vilab/In-Context-LoRA)
  •   - 确保生成图片的风格一致性
  •   - 生成关键剧情帧段
  • 3. **最终输出**:
  •   - 生成电影海报
  •   - 输出关键场景剧照
  • ## 部署选项
  • - **云端部署**:
  •  - 默认使用StepFun平台的云节点
  •  - 无需本地配置
  •  
  • - **本地部署**:
  •  - 可根据实际硬件配置替换为本地节点
  •  - 支持混合部署模式
  • ## 工作流程
  • 1. 用户上传角色信息
  • 2. 输入场景描述和故事概要
  • 3. 系统自动生成首张角色图像
  • 4. 通过LoRA技术生成连续性场景
  • 5. 最终合成电影海报
  • ## 优势特点
  • - 全云端操作,零本地依赖
  • - 模块化设计,支持自定义扩展
  • - 基于Flux的稳定工作流
  • - 高度可定制的部署方案
  • Discussion

    (No comments yet)

    Loading...

    Author

    1
    1.2K
    18
    4.6K

    No reviews yet

    • - latest (a year ago)

    Primitive Nodes (67)

    Anything Everywhere (4)

    BizyAir_ApplyPulidFlux (1)

    BizyAir_BasicGuider (3)

    BizyAir_BasicScheduler (2)

    BizyAir_CLIPTextEncodeFlux (3)

    BizyAir_DualCLIPLoader (1)

    BizyAir_KSamplerSelect (2)

    BizyAir_LoraLoader (1)

    BizyAir_PulidFluxEvaClipLoader (1)

    BizyAir_PulidFluxInsightFaceLoader (1)

    BizyAir_PulidFluxModelLoader (1)

    BizyAir_RandomNoise (2)

    BizyAir_SamplerCustomAdvanced (3)

    BizyAir_UNETLoader (1)

    BizyAir_VAEDecode (4)

    BizyAir_VAEEncode (1)

    BizyAir_VAELoader (1)

    CombineStrings (7)

    GetNode (11)

    Note (1)

    Reroute (1)

    SetNode (8)

    StepFunClient (1)

    TTP_Expand_And_Mask (1)

    TextImageChat (5)

    Custom Nodes (19)

    ComfyUI

    • - SetLatentNoiseMask (1)

    • - PreviewImage (3)

    • - EmptyLatentImage (2)

    • - SaveImage (2)

    • - LoadImage (1)

    pythongosssss/ComfyUI-Custom-Scripts

    • - ShowText|pysssss (8)

    • - Text Multiline (2)

    Checkpoints (0)

    LoRAs (1)

    filmstoryboard/film-storyboard.safetensors