sylm

In the workshop "From heavy load to lightning fast", Shu Zhao, CMO at @wavespeed_ai, broke down how to handle over 2M daily AI inferences, without losing speed. Upscale Conf is an event for all creators, technologists, and innovators. It's where bold creative minds connect to explore, debate, and show how technology can be used to achieve digital masterpieces.

In this repository, we present Wan2.1, a comprehensive and open suite of video foundation models that pushes the boundaries of video generation. Wan2.1 offers these key features:

SOTA Performance: Wan2.1 consistently outperforms existing open-source models and state-of-the-art commercial solutions across multiple benchmarks. Supports Consumer-grade GPUs: The T2V-1.3B model requires only 8.19 GB VRAM, making it compatible with almost all consumer-grade GPUs. It can generate a 5-second 480P video on an RTX 4090 in about 4 minutes (without optimization techniques like quantization). Its performance is even comparable to some closed-source models. Multiple Tasks: Wan2.1 excels in Text-to-Video, Image-to-Video, Video Editing, Text-to-Image, and Video-to-Audio, advancing the field of video generation. Visual Text Generation: Wan2.1 is the first video model capable of generating both Chinese and English text, featuring robust text generation that enhances its practical applications. Powerful Video VAE: Wan-VAE delivers exceptional efficiency and performance, encoding and decoding 1080P videos of any length while preserving temporal information, making it an ideal foundation for video and image generation.

Wan is an open source video generation modeling suite developed by Alibaba that supports a variety of tasks, including Text-to-Video, Image-to-Video, Video Editing, Text-to-Image, and Video-to-Audio.

With Wan-Trainer provided by WaveSpeedAI, you can easily upload your own dataset and quickly fine-tune your own Wan-LoRA model in a simple operation! All tasks supported by Wan — from Text-to-Video — can be trained directly on WaveSpeedAI.

Flux Dev is a non-commercial open source version of the text-to-image model developed by Black Forest Labs. It excels in image quality, detail performance and cue word reproduction. With Flux-Dev-LoRA-Trainer provided by WaveSpeedAI, you can easily upload your own dataset and fine-tune your own Flux-Dev-LoRA model in a simple operation!

HunyuanVideo-I2V is an open-source AI video generation model developed by Tencent, designed for image-to-video tasks. It offers two versions: a 14-billion-parameter professional model that excels in generating complex motions and simulating physical dynamics, and a 1.3-billion-parameter lite version optimized for consumer-grade GPUs, requiring only 8.2GB of VRAM, making it suitable for secondary development and academic research.

Built upon a causal 3D Variational Autoencoder (VAE) and Video Diffusion Transformer architecture, HunyuanVideo-I2V efficiently models spatiotemporal dependencies. In the authoritative VBench evaluation, the 14B version achieved a leading score of 86.22%, surpassing models like Sora, Luma, and Pika, and securing the top position. The model is available on Wavespeed AI, providing convenient access for developers.

Thank you very much for your suggestion. It's very useful to me.