VACE-14B Video Generation

Discover VACE-14B, your ultimate solution for AI-powered video generation and editing. Ready to experience the power of AI? Start your journey here!

Platform: Replicate

Video GenerationVideo EditingAll-in-one Video CreationPrompt-Based Editing

6.5k runs

H100

Commercial

🚀Function Overview

A model for generating and editing videos using text prompts, input videos, masks, and reference images with customizable parameters.

Key Features

Video generation from text prompts
Video editing with source video and mask inputs
Resolution and frame control
Speed optimization levels
Guidance scale and sample parameter adjustments

Use Cases

•Creating videos from textual descriptions
•Editing existing videos with masks or reference images
•Generating dynamic visual content
•Producing optimized video outputs for various resolutions

⚙️Input Parameters

prompt

string

Prompt

src_video

string

Input video to edit

src_mask

string

Input mask video or image to edit

src_ref_images

array

Input reference images to edit

speed_mode

string

Speed optimization level

frame_num

integer

Number of frames to generate

size

string

Output resolution

seed

integer

Random seed (-1 for random)

sample_shift

integer

Sample shift

sample_solver

string

Sample solver

sample_steps

integer

Sample steps

sample_guide_scale

number

Sample guide scale

💡Usage Examples

Example 1

Input Parameters

{
  "seed": -1,
  "size": "832*480",
  "prompt": "An apple being cut by a knife",
  "frame_num": 81,
  "speed_mode": "Lightly Juiced 🍊 (more consistent)",
  "sample_shift": 16,
  "sample_steps": 50,
  "sample_solver": "unipc",
  "sample_guide_scale": 5
}

Output Results

https://replicate.delivery/xezq/IKSvf6HkSXwIXKkf7la36IGefzsK5Z6abxzpnuDgtedvbCqlC/output.mp4

Quick Actions

Use NowView Documentation

Technical Specifications

Hardware Type: H100
Run Count: 6.5k
Commercial Use: Supported
Platform: Replicate

Related Keywords

Video CreationVideo EditingPrompt-Based EditingText-to-VideoMask-Based EditingHigh-Resolution OutputOptimized Performance

Related Models

twha Video Generation Model

A model for generating videos from text prompts and starting images.

Lightricks LTX Video Generation

DiT-based 13b video generation model, creating 30fps video

PixVerse V3.5 Video Generation

Create videos in as little as 10 seconds. 5s or 8s videos at 360p, 540p, 720p or 1080p.