Internet Watch
Luma: A large model for video generation of cinematic quality videos
Another new video generation model - Luma. It is said to be able to generate movie-quality videos. What is Luma AI? Luma AI is […]
Kling AI: Kuaishou's video generation model, comparable to Sora
Kling AI is a large video generation model released by Kuaishou, which can generate videos up to 2 minutes long. Its main advantage is that it uses Kuaishou's self-developed 3D expression and body reconstruction technology to drive expressions and body movements through a full-body photo of a person. Kling AI is suitable for generating scenes such as singing and dancing videos and long videos.
ChatTTS: A text-to-speech model for conversational scenarios
Overview Recently, a text-to-speech model has become popular, that is: ChatTTS. Moreover, this model was developed by a small team in China. Focusing on […]
GPT-4o: A free large mixed-input and mixed-output model
The release of GPT-4o marks a new milestone for large hybrid input-output models with richer and faster conversations. Its free use and reduced API fees make this technology accessible to more people. Compared to Gemini1.5 Pro, GPT-4o is superior in conversational experience and promotes the development of robots and AIGC applications. This progress represents an important development in the field of generative AI, bringing new possibilities to areas such as human-computer interaction and content generation.
Viggle AI: How to generate videos with controllable character movements
Video generation models such as Sora and Stable Video Dissfusion often face the problem of being unable to accurately control the output video, especially in terms of character movements. The controllable video model can accurately control the character movements in the video through prompt words. Viggle AI, as the first video-3D model with actual physical understanding capabilities, can freely control character movements and is embedded in the Discord platform. This controllable video technology will significantly reduce the cost of digital human products and enable diversified digital human video creation.
Google Gemini 1.5 Pro personal test: powerful and fragile at the same time
After testing the newly upgraded multi-modal AI model Gemini 1.5 Pro, users found that although it supports a more comprehensive input type including text, pictures, videos, files and folders, the reasoning ability has not been significantly improved, especially in distinguishing right from wrong. Additionally, processing of video, file, and folder inputs takes a long time, and there are limitations in handling large amounts of data.
Hot topics in February 2024: Sora - Open AI's large video generation model
On February 16, 2024, Open AI released its advanced video generation model named Sora, sparking interest almost rivalling that of GPT. Sora, which is not yet available for public use, combines Transformer and diffusion architectures for high-fidelity video simulation. Open AI's TikTok showcases Sora's capabilities with unedited videos from various prompts, previewing its potential impact in the burgeoning video generation field.
Gemini 1.5 pro: How to apply
Google Gemini1.5 pro overview Google Gemini1.5 pro on February 15, 2024 […]
Hot topics in January 2024: palworld
1. Google Trends: Compare “AI”, “gpt”, “palworld” This is a screenshot from today (2024/01/31). […]
The new WordPress experience: building websites with SAAS, low-code and no-code
On November 6, 2023, WordPress v6.4.2 was released. Two days later, I migrated my blog to another server. Later […]
