Alibaba Cloud Opens AI Video Generation Model 'Wan2.1'

Alibaba Cloud announced on the 27th that it will open source and release its artificial intelligence (AI)-based video generation model 'Wan2.1' for free.



The Wan2.1 series recorded a VBench LeaderBoss composite score of 86.22%. [Photo = Alibaba Cloud]



Alibaba Cloud is releasing four 14 billion (14B) and 1.3 billion (1.3B) parameter models from the Wan2.1 series, the latest version of its video foundation model 'Tongyi Wanxiang', as open source, further enhancing the openness and scalability of AI technology. The four



models released this time are T2V-14B, T2V-1.3B, I2V-14B-720P, and I2V-14B-480P, and are designed to generate high-quality images and videos based on text and image input. These models can be downloaded from Alibaba Cloud's AI model community, ModelScope, and collaborative AI platform, Hugging Face, and are open for free use by academic researchers and enterprises around the world.



The T2V-14B model is optimized for high-quality video generation with complex motions, while the T2V-1.3B model provides an ideal solution for a variety of developers conducting research and secondary development by balancing generation quality and computational efficiency. The I2V-

14B-720P and I2V-14B-480P models support not only text-based video generation but also image-based video generation functions.



The Wan2.1 series, released earlier this year, is the first AI video generation model to support text effects in Chinese and English. It excels in realistic video generation capabilities such as precise processing of complex motions, improved pixel quality, compliance with physical principles, and optimized command execution accuracy.



Wan2.1 ranked first on VBench Leaderboard, a comprehensive benchmark for video generation models. It is the only open source video generation model among the top five models on HuggingFace's VBench leaderboard.



According to VBench, the Wan2.1 series achieved a comprehensive score of 86.22%, demonstrating top-level performance in key evaluation items such as naturalness of movement, spatial relationship, color expression, and multi-object interaction.



Alibaba Cloud explained that "training video generation AI models requires massive computing resources and a large amount of high-quality training data," and "open sourcing these models can lower the barriers to AI utilization, and enterprises can produce high-quality video content optimized for their businesses in a more efficient and economical way."





https://www.inews24.com/view/blogger/1818307

Comments

Post a Comment

Popular posts from this blog

Livestock Manure Methane Is Soaring, But ‘Resource Recovery’ Isn’t Working [Now is a Climate Crisis]

Image
It has been pointed out that the methane emissions from the livestock waste treatment process were actually reported to be 2.5 times smaller than the previous calculation standards. It is expected that livestock waste will increase rapidly in the future, but an analysis report has been released that 'biogasification', which turns it into a resource, is only at the 10% level. It is urgent to come up with countermeasures as methane, one of the greenhouse gases, is expected to increase rapidly. According to the analysis of the entire livestock waste treatment process for the first time in Korea, the largest source of methane emissions was pigs (73%), and only 14% of pig waste was used for biogas in the public sector. It has been pointed out that it is urgent to come up with countermeasures as methane from livestock waste is increasing rapidly but 'resource conversion' is insufficient. [Photo = Climate Solution] About half of domestic methane...
Read more

KT-MS, 'Super Cooperation' to Improve Korea's AI Capabilities... "Providing AI Education to All Citizens" (Comprehensive)

Image
KT and Microsoft (MS) are going beyond corporate partnerships to engage in hyper-collaboration to enhance Korea's artificial intelligence (A)I capabilities. The plan is to provide AI education to all citizens interested in AI and contribute to improving the level of "national AI literacy" (the ability to understand and utilize AI technology). KT CEO Kim Young-seop (left) and Microsoft CEO and Chairman of the Board Satya Nadella pose for a commemorative photo after meeting at KT's Gwanghwamun office on the 25th. [Photo = KT] First official meeting of the two companies' CEOs in Korea···Discussion of plans to promote AX in Korea On the 26th, KT announced that KT CEO Kim Young-seop and Microsoft CEO and Chairman of the Board Satya Nadella met and exchanged opinions on accelerating the industry's AX (AI transition). The explanation is that they discussed agenda items that will help strengthen the country's AI competitivenes...
Read more

"TSMC to have 2nm line with monthly production capacity of 50,000 sheets by year-end"

Image
There is a report that TSMC, the world's largest foundry, is expected to have a 2nm process line capable of producing around 50,000 sheets per month by the end of this year. TSMC [Photo = TSMC] Taiwanese economic media Commercial Times reported on the 24th (local time) that TSMC's 2nm process production capacity could reach 50,000 sheets per month by the end of the year, and could increase to 80,000 sheets once the Baoshan and Kaohsiung factories are fully operational. TSMC's Baoshan factory is scheduled to receive its first order for the second half of this year from April 1, and its Kaohsiung factory in southern Taiwan is also reported to hold an expansion ceremony on the 31st. Kuo Ming-chi, an analyst at TF International, claimed via X (formerly Twitter) that "TSMC's 2nm trial production yield had already reached 60-70% three months ago, and is now well above that." TSMC has already secured an overwhelming share of th...
Read more