In today's technology landscape, CLIP (Contrastive Language-Image Pre-training) is an important multimodal foundational model. It combines visual signals and text signals into a shared feature space using contrastive learning loss on a large-scale dataset of image-text pairs. As a retriever, CLIP supports various tasks such as zero-shot classification, detection, segmentation, and image-text retrieval. Meanwhile, as a feature extractor, it performs well in nearly all...

The rapid development of generative AI has raised the challenge of comprehensively evaluating its performance. A plethora of models have emerged, each showcasing increasingly impressive results. However, the pressing question remains: how do we assess the effectiveness of these text-to-image models? Traditional evaluation methods often rely on subjective visual assessments or simple metrics like CLIPScore, which frequently fail to capture the intricate details found in complex text prompts, such as relationships between objects and logical reasoning. This leads to inaccurate evaluation outcomes for many text-to-image models.

Recently, an open source multimodal AI model named Molmo has drawn widespread attention in the industry. This AI system, based on Qwen2-72B and leveraging OpenAI's CLIP as the visual processing engine, is challenging the dominance of traditional commercial models with its outstanding performance and innovative features. Molmo's standout characteristic is its efficient performance. Despite its relatively small size, it can compete with competitors that are ten times larger in processing capability. This 'small but exquisite' design philosophy not only enhances the model's

Aishi Technology has launched PixVerse V2, an AI video generation tool designed to unleash user creativity. Utilizing a Diffusion+Transformer architecture with a spatiotemporal attention mechanism, it accurately understands and generates multimodal information, supporting efficient training and optimization. Key features include enhanced spatial and temporal perception, precise text alignment, and the ability to produce smooth, coherent video seg....

DeepSeek has open-sourced the RWKV-CLIP model, a visual-language representation learner that combines the strengths of both Transformer and RNN architectures. The model has significantly enhanced performance on vision and language tasks by pre-training on image-text pairs sourced from web data, expanding its dataset with information obtained from various websites.To address the issue of noisy data and improve data quality, the research team introduced a diverse description generation framework.

Microsoft announced today that Microsoft Edge and other Chromium-based browsers will support copying and pasting SVG (Scalable Vector Graphics) files through the Async Clipboard API. This significant breakthrough will significantly enhance user experience and web development efficiency. SVG format has always been the preferred choice for web graphics due to its high-quality visual effects and the ability to scale losslessly. However, web applications face many challenges when dealing with SVG co

Clipfly is Fotor's one-stop AI long video platform, integrating various AI video generation, enhancement, and editing features to assist users throughout the entire video production process. The design philosophy of Clipfly is to simplify the video production workflow, allowing users to generate image frames based on prompts and convert them into video clips, ultimately producing a complete video. Clipfly combines open-source models such as Stable Diffusion to enhance the quality of video clips and meet the needs of creators. After six months of development, Clipfly...

Jasper has successfully completed the acquisition of Clipdrop, the AI image platform under Stability AI. Clipdrop offers powerful AI image editing tools that enterprise clients can access through the Jasper API. Stability AI sold Clipdrop to Jasper due to financial difficulties. This acquisition will enhance Jasper's market share in the AI image processing sector.

Douyin has officially launched the AI live stream clipping feature, which may lead to the elimination of third-party teams. The new feature supports automatic editing and publishing of live stream replays, allowing merchants to manually edit and generate high-quality short videos. Live stream clipping has become an important means for top Douyin anchors to drive sales, but as more participants join, the problem of internal competition has become increasingly severe. The introduction of Douyin's official AI tools accelerates the elimination of third-party teams, intensifying competition. The clipping industry is massive, with top anchors generating revenues of several hundred million yuan from clipping-based sales, but issues related to resource competition and traffic also arise.