John Janiczek’s Post

Google's Gemini 1.5 Pro Al can now process audio files; outshines Ultra Google has enhanced its Gemini 1.5 Pro Al model, enabling it to process audio files from sources like earnings calls and videos without transcripts. Now available for public preview on the Vertex Al platform, the updated model has surpassed Gemini Ultra's performance. Google has also improved its Imagen 2 text-to-image generation model and introduced SynthID digital watermarking for image provenance.. . . . . . #gemini#Ai#techonolgy#digital#informative #coder#programer#computer#google#googleupdates

https://lnkd.in/g8C3EyTi Hello linkedin Professionals comunity! here's an overview of my new project file on" Quantum Virtual Machine (QVM)": It's a simulator that mimics Google's actual quantum hardware processors. You can run quantum circuits on the QVM just like you would on a physical quantum computer. The QVM uses simulations that incorporate noise data to accurately represent the errors and limitations of real Google quantum processors. Tests have shown the QVM's outputs to be very close to those of the real hardware. The QVM also employs a high-performance simulator called qsim to handle larger circuits efficiently. Please click below link to see my project files on QVM

https://lnkd.in/g8C3EyTi Hello linkedin Professionals comunity! here's an overview of my new project file on" Quantum Virtual Machine (QVM)": It's a simulator that mimics Google's actual quantum hardware processors. You can run quantum circuits on the QVM just like you would on a physical quantum computer. The QVM uses simulations that incorporate noise data to accurately represent the errors and limitations of real Google quantum processors. Tests have shown the QVM's outputs to be very close to those of the real hardware. The QVM also employs a high-performance simulator called qsim to handle larger circuits efficiently. Please click below link to see my project files on QVM

I've noticed many struggle with coding a simple PID controller and commonly encounter a few key mistakes: 1. Undefined Sampling Time: The sampling time must be explicitly defined, rather than letting the program run recursively with numbers that keep increasing. 2. Improper Coding of Differentiation and Integration: These operations need to be coded correctly for the PID controller to function as intended. 3. Unrealistic Motor Model: Often, the motor model used in simulations doesn’t reflect real-world behavior, leading to a gap between simulated tuning and what works with an actual motor. To address these challenges, I've coded a PID-based speed controller for a DC motor, but with a twist: the current settings make the motor unstable on purpose. This setup is a great exercise for those looking to practice and improve. Just copy my code into Google Colab to try it out. Your goal is to achieve stable output, with acceptable rise time, no oscillations, and no steady-state error. It's a practical way to get hands-on experience with tuning a PID controller. Note, that I've left out actuator saturation considerations for simplicity, but including it could lead to more realistic outcomes. Challenge yourself to adjust the parameters and add saturation handling to enhance the controller's realism and performance. #PIDController #ControlSystems #MotorControl #Simulation https://lnkd.in/dMk_fMjk

 🚀 Join us for the 4th DeFiChain Labs Engineering Sync!   📅 Next week, Monday, July 15, 2024 ⏰ Live at 11:30 (CET) // 17:30 (SGT) 🔗 Join here (https://lnkd.in/dm-5XP9G) 🕒 Duration: 45 minutes This call covers:   ✅ Status Updates ✅ Current Team Projects ✅ Collaborative Issue/Bug Squashing on GitHub ✅ Q&A Note: This is a technical session focused on engineering topics only, with no discussions on tokenomics, economics, or prices. Don’t miss this chance to stay informed and get involved with the DeFiChain Labs Engineering team!

Am I the first person to build an LLM model in Google Colab? Introducing, BartPhi-1.0! It currently utilizes Bart as the initial input layer, and then feeds that output to Phi-1.5. Phi-1.5 then outputs worse and less coherent text than it otherwise would as a result! It's a start though. I proved it can be done. No GPU. Now to play around with fine tuning BartPhi and different configurations to turn it into something that is actually useful!

🚀Join us today, for the 4th DeFiChain Labs Engineering Sync!   📅 Today, Monday, July 15, 2024 ⏰ Live at 11:30 (CET) // 17:30 (SGT) 🔗 Join here: https://lnkd.in/dm-5XP9G 🕦 Duration: 45 minutes This call covers: ✅ Status Updates ✅ Current Team Projects ✅ Collaborative Issue/Bug Squashing on GitHub ✅ Q&A Note: This is a technical session focused on engineering topics only, with no discussions on tokenomics, economics, or prices. Don’t miss this chance to stay informed and get involved with the DeFiChain Labs Engineering team!

Buckle up, video creators... Google VideoPoet is the next big thing waiting to happen in the video creation. What is it?.... Google VideoPoet, currently in the research phase, is a groundbreaking large language model (LLM) for zero shot video creation. What will it offer? - Text to Video : just requires a prompt like text to image generation - Image to Video -Video to audio - by default, VideoPoet outputs 2-second videos. But the model is also capable of long video generation by predicting 1 second of video output given an input of a 1-second video clip. Visit Google Research for more: https://lnkd.in/dsyFu59y The model is currently under development and not available publicly, but it has already produced some magic video graphics. #google #googlebard #googlevideopoet #videopoet #linkedin #googlenews

In complete agreement. The use of slang often reflects the values and attitudes of a particular generation, so if Gen Z finds inspiration and meaning in this phrase, then it is a testament to the versatility and adaptability of language when it comes to expressing complex ideas in an eloquent and relatable manner.

Google to Open First Code Next Lab at Hollywood Park MXU in Southern California Read the full article below..