Wan Flf2v

Wan Flf2v

Playground

Wan-2.1 flf2v generates dynamic videos by intelligently bridging a given first frame to a desired end frame through smooth, coherent motion sequences.

Features

Wan FLF2V

Wan FLF2V (First-Last Frame Video Generation) is an open-source video generation model developed by the Alibaba Tongyi Wanxiang team. Its open-source license is Apache 2.0. Users only need to provide two images as the starting and ending frames, and the model automatically generates intermediate transition frames, outputting a logically coherent and naturally flowing 720p high-definition video.

Core Technical Highlights

• Precise First-Last Frame Control: The matching rate of first and last frames reaches 98%, defining video boundaries through starting and ending scenes, intelligently filling intermediate dynamic changes to achieve scene transitions and object morphing effects.
• Stable and Smooth Video Generation: Using CLIP semantic features and cross-attention mechanisms, the video jitter rate is reduced by 37% compared to similar models, ensuring natural and smooth transitions.
• Multi-functional Creative Capabilities: Supports dynamic embedding of Chinese and English subtitles, generation of anime/realistic/fantasy and other styles, adapting to different creative needs.
• 720p HD Output: Directly generates 1280×720 resolution videos without post-processing, suitable for social media and commercial applications.
• Open-source Ecosystem Support: Model weights, code, and training framework are fully open-sourced, supporting deployment on mainstream AI platforms.

Technical Principles and Architecture

• DiT Architecture: Based on diffusion models and Diffusion Transformer architecture, combined with Full Attention mechanism to optimize spatiotemporal dependency modeling, ensuring video coherence.
• 3D Causal Variational Encoder: Wan-VAE technology compresses HD frames to 1/128 size while retaining subtle dynamic details, significantly reducing memory requirements.
• Three-stage Training Strategy: Starting from 480P resolution pre-training, gradually upgrading to 720P, balancing generation quality and computational efficiency through phased optimization.

# Submit the task
curl --location --request POST "https://api.wavespeed.ai/api/v3/wavespeed-ai/wan-flf2v" \
--header "Content-Type: application/json" \
--header "Authorization: Bearer ${WAVESPEED_API_KEY}" \
--data-raw '{
    "first_image": "https://raw.githubusercontent.com/Comfy-Org/example_workflows/refs/heads/main/wan2.1_flf2v/input/start_image.png",
    "last_image": "https://raw.githubusercontent.com/Comfy-Org/example_workflows/refs/heads/main/wan2.1_flf2v/input/end_image.png",
    "prompt": "glass flower blossom",
    "negative_prompt": "",
    "duration": 5,
    "size": "832*480",
    "num_inference_steps": 30,
    "guidance_scale": 5,
    "seed": -1,
    "enable_safety_checker": true
}'

# Get the result
curl --location --request GET "https://api.wavespeed.ai/api/v3/predictions/${requestId}/result" \
--header "Authorization: Bearer ${WAVESPEED_API_KEY}"