Groundlight retweeted
We're excited to invite you to an exclusive webinar on Counting Objects with Computer Vision, featuring @GroundlightAI!
Join Dr. Paulina Varshavskaya and Dr. Amirreza Farnoosh F. as they discuss how you can:
✅ Count the exact things you care about - for example, go beyond counting the number of people and count the number of people in your company's uniform.
✅ Combine active learning with real-time human oversight so you can configure it to perform at any accuracy level you want.
💻 Register Here: app.livestorm.co/groundlight…
📅 Date: Thursday, April 10th
⏰ Time: 12:00 PM - 12:45 PM PST
#AI #MachineLearning #ComputerVision #Automation #Manufacturing #DigitalTransformation
Groundlight retweeted
We trained a Visual LLM to reason using GRPO, and open sourced the code. Tiny 3B model beats all the big players (GPT, Claude, etc 0-shot) after RL training on this cryptogram task. Live demo and links: groundlight.ai/blog/visual-r…
Vision foundation models have kinda stagnated recently, but now that we have shown how to incorporate reason, I think we'll be able to make progress again.
Groundlight retweeted
Wow! I'm super excited for Groundlight to be nominated for GeekWire's AI Innovation of the Year. It's an honor to be listed in such a fantastic group. Pop on over and give us a vote plz?
geekwire.com/votenow
(AI is the new category at the bottom.)
Groundlight retweeted
Here's the story ... geekwire.com/2024/how-this-s…
Groundlight retweeted
Excited to be at @Robo_Business this week. Want to talk about adding simple reliable perception to your robotic system? DM me.
Groundlight retweeted
Sharing one more time before we go live tomorrow morning! I'll be talking with @leopd from @GroundlightAI about computer vision and robotics on this months @union_ai fireside chat 🔥
eventbrite.com/e/the-future-…
Groundlight retweeted
If you were ever curious about how active learning can reduce data labeling costs for computer vision and machine learning, here's a blog post I wrote about how we do it at @GroundlightAI
code.groundlight.ai/python-s…
Groundlight retweeted
Say you get 94% accuracy on a k-fold eval of your model, and that this number is very important to you. K-fold is just an estimate of the "true" accuracy, but how good an estimate? How might you calculate a confidence interval? Here's the math: code.groundlight.ai/python-s…





