Unverified50% confidenceFactExact time
VAE decoding may account for 30% to 60% of the overall video generation time.
1
Sources
50%
Confidence
Long-term
Relevance
9/24/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn image-to-video generation, the VAE (Variational Autoencoder) acts as an image compressor, with a typical compression ratio of 8×, meaning a 720p image is encoded into a 90×160 feature map.71% similarUnverifiedThe VAE (Variational Autoencoder) is responsible for encoding and decoding between latent space and pixel space, and is one of the significant VRAM consumption points in video/image generation pipelines.71% similarUnverifiedProfessional YouTubers typically fill 60%–70% of their video runtime with B-roll.67% similarUnverified视频生成任务的显存占用约为同分辨率图像任务的10–30倍67% similarUnverifiedIn long-video scenarios exceeding 20 minutes, the CaptionQA approach using 30-second and 60-second windows for caption generation outperformed the conventional direct VideoQA approach.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/945722API
curl https://kongchang.com/api/v1/knowledge/claims/945722MCP
get_claim(id=945722)