You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
You now can train a really good image classification model on 25k images in less than 2 hours, entirely for free, on a Colab GPU. And it's dead easy. What a time to be alive I remember when you had to write your own CUDA and C++ and you needed a hardware budget
Here's a Colab that does just that. Download a zip file with 25k images. Get a model training in seconds, with all the best practices baked in. One hour and a half to convergence on the free Colab GPU.https://colab.research.google.com/drive/1umJnCp8tZ7UDTYSQsuWdKRhqbHts38AC …
Have you done the same on edge eg with google Coral?
Indeed, the infrastructure + software have improved a lot. Thanks to all those that have contributed! 
I just finished a TensorFlow in Practice Coursera specialization and was surprised by what you described. I am relatively new to the world of ML.
yes how about when you had to write your own distributed least squares optimization as well?
I love Colab. No setups at all. Google rocks.
Assuming this is just training the top n layers of a pre-trained network (VGG etc.)? Can’t imagine training a general purpose image classification DNN with just 25,000 examples.
If I recall correctly a chap from Walmart labs put the figure at 10k for what they were talking about at a data festival - admittedly, they weren’t going to suffer in the same way from misclassification as other applications so probably had higher tolerance.
Is it worth learning cuda and c++ in this day and age. Are there any practical reasons you could come up with ?
Well, all deep learning learning runs on C++ and CUDA. Python is just the user API for accessing all the functionality.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
