BEIJING, June 11, 2024 (GLOBE NEWSWIRE) -- Cheer Holding, Inc. (NASDAQ: CHR) ("Cheer Holding," or the "Company"), a leading provider of advanced mobile internet infrastructure and platform services, today announced the official release of the upgraded 2.0 edition of its multimodal AI creative platform CHEERS Telepathy (the “Platform”). This edition features more advanced and complicated algorithms and models, more powerful application capacity, and more comprehensive AI interaction functionalities. This marks a breakthrough in the underlying technology of the Platform, and a qualitative leap at the application level, bringing an unprecedented user experience, as well as a richer, more diverse and authentic generation effect.

In this grand upgrade, The Platform launched a brand new “text-to-video” tool, which is fully accessible to all users. Benefiting from the following technological innovations and applications, the “text-to-video” technology goes far above the industry average level.

1. DiT hybrid model architecture: The underlying technology of the Platform adopts the cutting-edge DiT (Diffusion Transformer) hybrid model architecture, which is different from the diffusion model of the UNet architecture widely used in the current AI video field. The Platform is able to generate high-quality video content leveraging the advantages of both the Diffusion model and the Transformer model. Specifically, the Diffusion model can add noises to the video frame through the diffusion process, and then de-noise the video frame gradually through the reverse diffusion process to generate the high-quality video frame with rich details; the Transformer model, with its self-attention mechanism, can capture and generate complex dependencies in the frame sequence to ensure the temporal consistency and coherence of the video frames.
   
   

2. In-depth simulation and controllability: The Platform can not only generate high-quality images, but also simulate the real physical world in a profound way. Assisted by DRL deep reinforcement learning, the Platform is able to understand and simulate complex real-world scenes and the moving of objects, providing users with more realistic and accurate video content. Meanwhile, this technology is also highly controllable, allowing users to adjust and optimize the generated videos based on their own needs to achieve personalized and creative expressions.
   
   

3. Unique creativity: The Platform can creatively generate video content of different styles based on the user's text description instructions. Empowered by the CGV algorithm in the model, it is able to capture and understand the user's creation intentions and transform them into visual video content.
   
   

4. Semantic understanding and localization advantages: Relying on the powerful support of the large model of Polaris Intelligent Cloud, The Platform performs well in semantic understanding, especially in the understanding under the Chinese context. This makes it easy for users to express both complex emotions and delicate cultural elements when creating video contents.

The Platform's underlying technology is advanced both in its innovative model architecture and powerful processing power, and its capacity to well understand users' demands and easily adapt to local culture. The integration and application of these technologies make the Platform an efficient, accurate and creative video generation tool for users.

The product team not only greatly optimized its core functions, but also added the innovative AI Image Upscaler and the AI avatar production tool, further expanding the application of the Platform in commercial scenarios, and providing more diversified marketing tools for commercial brands.

Utilizing advanced AI technology, the AI Image Upscaler can intelligently identify and upscale images, helping users easily achieve amplification and creative expansion of high-quality images without losing textures or details. Whether it is product detail display or high-definition presentation of advertising images, this tool can help meet high-standard business needs. The AI avatar production tool allows users to quickly generate unique avatars through personalized algorithms, providing a new way of interaction that will benefit social media marketing, virtual assistants and personalized services.

In addition, the Platform added the AI course module which is loaded with a large number of carefully filmed AI courses, aiming to facilitate the quick understanding and use by ordinary users. This provides strong service and technical support for individual creators.

With the continuous iterative upgrade of the Platform and the great improvement of its capabilities, the Company further demonstrated the strong technical R&D strength and product innovation capability, as well as its great development potential in the huge AI application market.