Using a source image, we prompted the two models with, “Replace the background with the interior of a mall.” (2/6)
MULTIMODAL AI
-
FLUX Kontext Pro and GPT-Image-1 Enable Conversational Image Editing
By
–
With FLUX Kontext Pro and GPT-Image-1, you can make conversational edits to images while preserving the subject. This includes use cases like altering appearances, adding objects, and/or replacing backgrounds. Here are a few examples comparing the two models on Poe
-
Veo 3 Generates Unhinged Flight Safety Video With Googly Eyes
By
–
I asked for an unhinged flight safety video, and Veo 3 put googly eyes on the oxygen mask.
-
Veo 3 Fast Released at One-Fifth Regular Pricing
By
–
This was with Veo 3 fast, so 1/5 the price of regular Veo
-
Reading Recognition Dataset from Egocentric Wearable Sensors
By
–
Reading Recognition in the Wild: A Dataset for Understanding Human Behaviors During Reading from an Egocentric Sensor Suite This large multimodal dataset features video, eye gaze, and head pose sensor outputs, created to help solve the task of reading recognition from wearable
-
AI Makes Character Animation Easier with Open Source Models
By
–
De repente animar personajes se vuelve mucho más fácil con la IA🔥 Esto se consigue a partir de una imagen plana, y usando todos los avances que modelos opensource como Wan van incorporando. Súper útil! https://t.co/E7YPL7rTzv
— Carlos Santana (@DotCSV) 13 juin 2025De repente animar personajes se vuelve mucho más fácil con la IA Esto se consigue a partir de una imagen plana, y usando todos los avances que modelos opensource como Wan van incorporando. Súper útil!
-
Computer Vision, 3D Reconstruction, and Robotics: One Integrated Space
By
–
Here a simulation -> deployment video
— Andreas Klinger 🦾 (@andreasklinger) 13 juin 2025
Imho doesn't show much tbh – but maybe it drives home why i keep repeating that computer vision, 3D reconstruction, simulation, spatial reasoning, simulation, robotics, even XR or gaussian splats, etc are all basically one big space.
If… pic.twitter.com/h97tR9ZPLXHere a simulation -> deployment video Imho doesn't show much tbh – but maybe it drives home why i keep repeating that computer vision, 3D reconstruction, simulation, spatial reasoning, simulation, robotics, even XR or gaussian splats, etc are all basically one big space. If
-
Stable Diffusion 3.5 Optimization Delivers 2.3x Faster Generation
By
–
As announced at @NVIDIAGTC today, we've collaborated to optimize Stable Diffusion 3.5 using TensorRT and FP8 quantization. Compared to the base @PyTorch models, these optimizations deliver: 2.3x faster generation with SD3.5 Large 1.7x faster generation with SD3.5 Medium
-
KerasHub: 70+ Pretrained Models Supporting JAX, TF, PyTorch
By
–
KerasHub is a collection of over 70 popular pretrained model architectures — LLMs, VLMs, image generation models, etc — that work with JAX, TF, PyTorch. They all support HuggingFace checkpoints — you can load any HF model with them for the corresponding architecture.
-
Copilot Vision Now Free on Windows with Real-Time Help
By
–
Copilot Vision on Windows is now generally available – and it's free!
— Mustafa Suleyman (@mustafasuleyman) 12 juin 2025
While you talk + screenshare Copilot can get you answers and help you navigate challenges real-time. For example, Copilot can “highlight" where to click to improve the lighting in your photo. pic.twitter.com/5l9zQXJr3rCopilot Vision on Windows is now generally available – and it's free! While you talk + screenshare Copilot can get you answers and help you navigate challenges real-time. For example, Copilot can “highlight" where to click to improve the lighting in your photo.