astroml / astroml-workshop_aas235

Workshop material for the AAS235 meeting

Machine Learning With AstroML

Workshop at the 235th Meeting of the AAS in Honolulu

• DATE: Monday, 6 January 2020
• TIME: 13:00 - 17:00pm
• LOCATION: ROOM 323 A at the Hawai'i Convention Center

PRE-WORKSHOP SETUP

Please be sure that you have a GitHub account.

Description

This workshop will introduce the astronomical community to the 2nd edition of the book Statistics, Data Mining, and Machine Learning in Astronomy and the associated software package astroML. The goal is to introduce participants to a variety of statistical and machine learning tools available within the open source astroML library. The format will be interactive, including short presentations on different machine learning methodologies followed by instructor-guided, Jupyter notebook based tutorials. In these tutorial sessions participants will be able to try out the tools and to ask questions from expert users and developers.

Our primary focus will be on the new material and applications in the 2nd edition of the book.

astroml-workshop_aas235's Issues

AAS2019_NNexample.ipynb plot_model_history(vgg_model.history) - KeyError: 'loss'

See attached screen print

below

Running sample code got:

KeyError: 'loss'

KeyError Traceback (most recent call last)
/var/folders/39/hmv9hpy10cg6cx015773lbtw0000gn/T/ipykernel_2037/1962927886.py in
1 # plot the training history
----> 2 plot_model_history(vgg_model.history)

/var/folders/39/hmv9hpy10cg6cx015773lbtw0000gn/T/ipykernel_2037/664928204.py in plot_model_history(history)
92
93 # Extract loss and accuracy
---> 94 loss = history.history['loss']
95 val_loss = history.history['val_loss']
96 acc = history.history['accuracy']

KeyError: 'loss'

Pre-workshop e-mail to participants

Nothing needs to be installed before the workshop, but we'll need participants to have a github account for login.

Best to ask them in a pre-workshop e-mail.