Vaibhav Srivastav’s Post

Microsoft just silently dropped Florence 2 👀Vision model that can tackle many vision tasks (captioning, long captioning, detection, region proposal, OCR) 🤏Small models (200M and 800M) with ~quality to models 100x larger 🎉MIT licensed Papers and models: https://lnkd.in/dTrA3vH4

Microsoft Research introduces VASA, a framework for generating lifelike talking faces of virtual characters with appealing visual affective skills (VAS), given a single static image and a speech audio clip. https://lnkd.in/gz3B8U66

Learn More about #ComputerVision #DataScience #Machineleaning:- PoseNet: Google's real-time machine learning model for human pose estimation, detecting key body parts in images or video frames. MS Transformer: Microsoft's extension of the Transformer architecture to computer vision tasks, leveraging self-attention for image classification, object detection, and semantic segmentation. Ours (Marepo): Without specific context, it's difficult to provide concise details, but "Ours (Marepo)" could refer to a proprietary model or project, potentially developed by a research group named Marepo, focusing on a specific area within machine learning or computer vision.

I'm excited to share my new article on Medium diving deep into the impressive Microsoft Florence 2 model! This tiny titan in computer vision packs a powerful punch, excelling in tasks like image classification, object detection, and even generating captions. Let me know in the comments what you think about Florence 2's potential! #Microsoft #MicrosoftFlorence2 #AI #ComputerVision #MachineLearning

Technical Deep Dive: Explore how federated learning enabled by NVIDIA FLARE can address data management challenges with easy and scalable integration to enhance the accuracy and robustness of your LLMs. https://nvda.ws/3TxVZUb

Without metadata, NO data can be found. Try this for real, our mind-blowing AI-IP story, based on fact. Is there any AI products you know, that can interpret and answer the following Chinese-English multilingual questions? "Who, in the Ontario province of Canada, has new US patents granted on the nearest Tuesday, when the USPTO releases the newly granted US patents on a weekly basis?" "Who, in the "江蘇‘’ province of China, has new US patents granted on the nearest Tuesday, when the USPTO releases the newly granted US patents on a weekly basis?" With our intellectual property (IP), a copyrighted multilingual metadata, we can answer. Metadata is an enabler. Without metadata, NO data can be found/retrieved, even by the most advanced technologies, like AI, NVIDIA chips, supercomputers, etc. https://lnkd.in/g-aJFnXR Experiment results showed that, with our intellectual property (IP), a copyrighted multilingual metadata, we are doing what AI, like ChatGPT, can't do in data analytics. Our IP can also make your information service UNIQUE in the world. Do you or any of your contacts need our expertise and our IP? We're selling, NOT just talking. " Thanks.

Hello connections, I want to share that one more my article on 🔉 medium.com The article is about key metrics to measure federated learning. Recently I have started writing one survey research paper on plant leaf 🌿 disease identification using federated learning with CNN,this topic is really interesting 🤔🧐. I started understanding of how FL works,I want to explain with simple example is that now a days everyone using smartphone that to every person have unique typing habits what FL will do it will predict the next typing word is going to be , initially every device is one local model for easy prediction after that aggregating all the local models to build one global model. Google uses FL to predict the next word.💫💫 https://lnkd.in/gR9z_9Jm

"Over the moon to announce my finishing touch of the 'Introduction to Machine Learning' direction with Great Learning! 🚀 Excited to discover the limitless opportunities of facts technology and device mastering. 🌟 #MachineLearning #DataScience #GreatLearningJourney"

#Microsoft introduced Florence-2, a cutting-edge vision foundation model that revolutionizes computer vision and vision-language tasks. Unlike traditional models, Florence-2 excels in understanding complex spatial hierarchies and semantic granularities with simple text prompts

Today I earned my "Fundamentals of Computer Vision" badge! I'm thrilled to share some exciting experiences with cloud technologies, particularly in building rapid, large-scale data applications. I've had the opportunity to explore Azure Vision Studio, creating some fascinating computer vision use cases: 🔘 Image Captioning: This innovative application aims to enhance customer experience in expansive retail environments. By providing descriptive captions for images, it offers valuable assistance to shoppers navigating large stores. 🔘 Object Detection and Summarization in Videos: Focused on improving efficiency in industrial settings, this application identifies and summarizes objects on factory floors. It's a game-changer for streamlining operations and enhancing safety protocols. These projects showcase the incredible potential of cloud-based AI services in solving real-world challenges. I'm continually amazed by how these technologies can transform various industries and improve everyday experiences. Have you had any similar experiences with cloud technologies or computer vision? I'd love to hear your thoughts or discuss any exciting projects you've encountered!