Scaling Laws for Native Multimodal Models Scaling Laws for Native Multimodal Models Shukor et al.: https://
arxiv.org/abs/2504.07951 #ArtificialIntelligence #DeepLearning #MachineLearning
MULTIMODAL AI
-
Scaling Laws for Native Multimodal Models
By
–
-
Major update: ultra-realistic 4K photos in one click
By
–
I just finished developing a major update on Shoot AI! You can now generate ultra-realistic photos of yourself, with 4K details. No need to create a new AI model, just click a button! ( + the video is already available in beta for
-
Tenstorrent: The Platform for Running AI in Video, Speech, and Code
By
–
🚨 Mise à jour : WordPress vient de lancer un créateur de site web gratuit propulsé par l’intelligence artificielle.
— VISION IA (@vision_ia) 11 avril 2025
Ll vous suffit de lui dire ce que vous voulez, et il construit automatiquement un site pour vous.
Voici tout ce que vous devez savoir : pic.twitter.com/xVd7pc86CfMise à jour : WordPress vient de lancer un créateur de site web gratuit propulsé par l’intelligence artificielle. Ll vous suffit de lui dire ce que vous voulez, et il construit automatiquement un site pour vous. Voici tout ce que vous devez savoir :
-
Google begins limited rollout of Veo 2 on AI Studio
By
–
BREAKING : Seeing multiple reports that Veo 2 is rolling out on Google AI Studio. Seems like it is a very limited rollout for now. h/t @HarisHiew
-
DDT: Decoupled Diffusion Transformer for Image Generation
By
–
DDT: Decoupled Diffusion Transformer
Hugging Face：
https://
huggingface.co/papers/2504.05
741
…
Paper:
https://
arxiv.org/pdf/2504.05741
Code：
https://
github.com/MCG-NJU/DDT -
Scaling Laws for Native Multimodal Models
By
–
Paper：Scaling Laws for Native Multimodal Models
Link: https://
arxiv.org/pdf/2504.07951 -
Apple Researchers Discover Scaling Laws Native Multimodal Models
By
–
Researchers from Apple just uncovered scaling laws for native multimodal models. What did they find?
– No clear advantage for late-fusion over early-fusion.
– Early-fusion performs better at small scales.
– Add MoEs, and you unlock modality-specific weights→big performance gains -
Groundbreaking video manipulation technology enables hyper-realistic character object editing
By
–
Introducing a groundbreaking new technology that allows you to hyper-realistically manipulate any character or object in your videos. These clips from our creator community are unreal…
— Pika (@pika_labs) 10 avril 2025
[1/7] @estebandiba pic.twitter.com/SWGi5bNM2CIntroducing a groundbreaking new technology that allows you to hyper-realistically manipulate any character or object in your videos. These clips from our creator community are unreal… [1/7] @estebandiba
-
Gemini 2.5 Pro Multimodal OCR Extracts Data from Dusty Records
By
–
Nice example of the multimodal OCR capabilities of Gemini 2.5 Pro. Just give it dusty old records and ask for JSON! https://
climatelabbook.substack.com/p/data-rescue-
with-ai
… "Which is almost perfect. It has misread a small number of the hard-to-read 8s as 5s. But it is spot-on apart from that. No -
AI System Now References Past Conversations for Contextual Responses
By
–
In addition to the saved memories that were there before, it can now reference your past chats to deliver responses that feel noticeably more relevant and useful. New conversations naturally build upon what it already knows about you, making interactions feel smoother and