Multimodal

Google expands its Gemini model family with the release of gemini embedding 2. This second generation model enables text-only success gemini-embedding-001 And it is specifically designed to address the high-dimensional …

Microsoft has released fi-4-logic-vision-15bA 15 Billion Parameter Open-Weight Multimodal Reasoning Model Designed for image and text tasks that require both perception and selective reasoning. It is a compact model designed …

How can a trillion-parameter large language model achieve state-of-the-art enterprise performance while reducing its total parameter count by 33.3% and increasing pre-training efficiency by 49%? Yuan Lab AI releases Yuan3.0 …

Google Research is proposing a new way of creating accessible software with Natively Adaptive Interfaces (NAI), an agentic framework where a multimodal AI agent becomes the primary user interface and …

section(“6) pack unpack”) B, Cemb = 2, 128 class_token = torch.randn(B, 1, Cemb, device=device) image_tokens = torch.randn(B, 196, Cemb, device=device) text_tokens = torch.randn(B, 32, Cemb, device=device) show_shape(“class_token”, class_token) show_shape(“image_tokens”, image_tokens) …

A team of Stanford Medicine researchers has introduced SleepFM Clinical, a multimodal sleep foundation model that learns from clinical polysomnography and predicts long-term disease risk from a single night of …

Meta Researchers Introduce Perception Encoder Audiovisual, PEAVAs a new family of encoders for joint audio and video understanding. The model learns aligned audio, video and text representations in a single …

Google has published T5Gemma 2open family encoder-decoder Custom-made transformer checkpoints Gemma 3 Pre-trained weights in an encoder-decoder layout, then continuing pre-training with that UL2 Objective. is released pre trained onlyThe …

Meta has released SAM Audio, a quick-driven audio separation model that targets a common editing constraint, separating a sound from a real-world mix without creating a custom model per sound …