CivArchive
    Qwen故事短剧双主体用指定背景2 - v1.0

    🎨 Discover Amazing AI Workflows on RunningHub!

    🎨 Discover Amazing AI Workflows on RunningHub!🔗 Click here to get this workflow: https://www.runninghub.ai/post/2018860398941839362/?inviteCode=rh-v1159

    🔗 Click here to get this workflow: https://www.runninghub.ai/post/2018860398941839362/?inviteCode=rh-v1159

    ### ComfyUI Workflow Introduction: Qwen Story Short Drama Dual Subject with Designated Background

    **Workflow Title:** Qwen Story Short Drama Dual Subject with Designated Background

    **Media Type:** Video (.mp4)

    **Number of Nodes:** 25

    #### Key Nodes and Their Functions:

    1. **CR Text**: This node is used for text conditioning to generate narrative elements for the story.

    2. **VAELoader**: Loads the VAE model necessary for encoding and decoding images, ensuring high fidelity in visuals.

    3. **VAEDecode**: Responsible for decoding the latent representations back into high-quality images while preserving original features.

    4. **SeedVR2LoadVAEModel**: Loads the VAE model with a specific seed for consistency across generated images, crucial for maintaining character features.

    5. **MarkdownNote**: Provides notes and documentation within the workflow for easy reference on node functionality and usage.

    6. **LayerUtility: PurgeVRAM V2**: Frees up VRAM by purging unnecessary data, optimizing performance during video generation.

    7. **LoadImage**: Imports the designated background images necessary for the story setting.

    8. **LoraLoaderModelOnly**: Loads only the model parameters for Lora, optimizing resource usage without loading full models.

    9. **EmptyLatentImage**: Initializes empty latent images for processing, essential for clean starts in image generation.

    10. **ConditioningZeroOut**: Resets conditioning inputs, allowing for fresh inputs in subsequent generations.

    #### Quick Usage Tips:

    - To **load images**: Use the **LoadImage** node to import your background images.

    - For **text generation**: Utilize the **CR Text** node to input narrative prompts.

    - Maintain **character consistency**: Ensure you set the correct seed using **SeedVR2LoadVAEModel**.

    - To manage resource efficiency, regularly use **LayerUtility: PurgeVRAM V2** to clear unnecessary data.

    - For **documentation**: Refer to the **MarkdownNote** for quick tips on using nodes effectively.

    #### Workflow Link:

    [Access the Full Workflow Here]()

    This workflow is designed to create compelling short dramas with a focus on maintaining the integrity of the original images while providing a cohesive storytelling experience. Happy creating! 🎥✨

    ### ComfyUI Workflow Introduction: Qwen Story Short Drama Dual Subject with Designated Background

    **Workflow Title:** Qwen Story Short Drama Dual Subject with Designated Background
    **Media Type:** Video (.mp4)
    **Number of Nodes:** 25

    #### Key Nodes and Their Functions:

    1. **CR Text**: This node is used for text conditioning to generate narrative elements for the story.

    2. **VAELoader**: Loads the VAE model necessary for encoding and decoding images, ensuring high fidelity in visuals.

    3. **VAEDecode**: Responsible for decoding the latent representations back into high-quality images while preserving original features.

    4. **SeedVR2LoadVAEModel**: Loads the VAE model with a specific seed for consistency across generated images, crucial for maintaining character features.

    5. **MarkdownNote**: Provides notes and documentation within the workflow for easy reference on node functionality and usage.

    6. **LayerUtility: PurgeVRAM V2**: Frees up VRAM by purging unnecessary data, optimizing performance during video generation.

    7. **LoadImage**: Imports the designated background images necessary for the story setting.

    8. **LoraLoaderModelOnly**: Loads only the model parameters for Lora, optimizing resource usage without loading full models.

    9. **EmptyLatentImage**: Initializes empty latent images for processing, essential for clean starts in image generation.

    10. **ConditioningZeroOut**: Resets conditioning inputs, allowing for fresh inputs in subsequent generations.

    #### Quick Usage Tips:

    - To **load images**: Use the **LoadImage** node to import your background images.
    - For **text generation**: Utilize the **CR Text** node to input narrative prompts.
    - Maintain **character consistency**: Ensure you set the correct seed using **SeedVR2LoadVAEModel**.
    - To manage resource efficiency, regularly use **LayerUtility: PurgeVRAM V2** to clear unnecessary data.
    - For **documentation**: Refer to the **MarkdownNote** for quick tips on using nodes effectively.

    #### Workflow Link:
    [Access the Full Workflow Here]()

    This workflow is designed to create compelling short dramas with a focus on maintaining the integrity of the original images while providing a cohesive storytelling experience. Happy creating! 🎥✨

    Description

    Workflows
    Qwen

    Details

    Downloads
    30
    Platform
    CivitAI
    Platform Status
    Deleted
    Created
    2/4/2026
    Updated
    4/27/2026
    Deleted
    2/6/2026

    Files

    qwen2_v10.zip

    Mirrors

    Huggingface (1 mirrors)
    CivitAI (1 mirrors)