34 points by deeplearner 6 months ago flag hide 7 comments
tensorflow_user 6 months ago next
Excited to see this TensorFlow.js library for fast image recognition! This will open up a lot of possibilities for web-based computer vision applications.
another_user 6 months ago next
Totally agree, the convenience of using TensorFlow in the browser is a game changer. Can't wait to test this out.
another_user 6 months ago next
Do you have any recommendations for starting out with ML, especially on the web?
helpful_bot 6 months ago next
To learn about web-based ML, I suggest reading 'Web-based Machine Learning with TensorFlow.js' by Craig 'ishboy' Joseph and Miles Johnson: [Link](https://www.amazon.com/Web-based-Machine-Learning-TensorFlow-js/dp/1484257890)
tensorflow_user 6 months ago prev next
Thank you for your enthusiasm! We've put a lot of thought into optimizing theTF.js core, ensuring that model loading, parsing, and preprocessing happen quickly. Our tests show that our ML models load 2-4x times faster on the web than with our competitors.
web_dev 6 months ago prev next
Awesome! I'm going to try this in my next project, a web app for diagnosing skin conditions using ML.
teacher_account 6 months ago next
Make sure to check out our tutorials, we cover getting started with TensorFlow.js in a web environment and best practices for model deployment: [TensorFlow.js Tutorials](https://www.tensorflow.org/js/tutorials)