The talk I gave was an updated version of the FOSSASIA workshop I gave previously, but without enough time for people to follow along with their own Virtual Machines. Also, because the audience mix was a little different, I chose to spend approximately equal time on four sections :
- Machine Learning in TensorFlow Playground
- Using Keras model zoo for the pre-trained NASNet model, and doing transfer learning there
- Going through the Reptile meta-learning algorithm (implemented in PyTorch)
- Showing how TensorFlow Eager mode can accomplish the same, line-by-line
As usual, all the source for the deeplearning-workshop notebooks is available on GitHub - if you have questions on the software, please leave an ‘issue’ there.
If there are any questions about the presentation please ask below, or contact me using the details given on the slides themselves.
- Presentation (50)
- ImageNet (4)
- TransferLearning (5)
- MetaLearning (3)
- OneShotLearning (3)
- Reptile (3)
- AIDay (2)
blog comments powered by Disqus