- atelierhide
- 5132
- 0
- 0
- 0
Training and testing crucial to effective machine learning. #scipy2015 #MachineLearning
2015-07-06 22:12:04"if someone says 'do you want more data? you say yes.’ “ more labeled data = better algorithms #scipy2015 #MachineLearning
2015-07-06 22:18:36wow, #cython turns a 2 line .pyx file into 16k lines of C. #SciPy2015
2015-07-06 22:25:25I fear that with all of of the Python experts out of the building this week (#scipy2015) we're going to revert to IDL and supermongo. #STScI
2015-07-06 22:29:09from #skimage at #scipy2015, re: jet: Is that a beauty mark on the man's upper lip? No, it's just an artifact of this ridiculous colormap.
2015-07-06 22:31:07"1000x speedup is not uncommon" -- diving into cython with Kurt Smith #SciPy2015
2015-07-06 22:41:03uncleaned comma data = the worst. #scipy2015 #cleanyourdata
2015-07-06 22:42:08having #cython for breakfast this morning, yum. #scipy2015
2015-07-06 22:43:09if someone else has extracted features you don’t have to be an expert in the domain that the data were collected #scipy2015 #MachineLearning
2015-07-06 22:47:39#swcscipy is getting started! #scipy2015 pic.twitter.com/TjQmzkIBFy
2015-07-06 22:48:03Look at your data before you try to do things with it. #scipy2015 #MachineLearning #alsolife
2015-07-06 22:56:17The Olivetti face data are SO CREEPY #SciPy2015 #MachineLearning pic.twitter.com/PLUTxQgxl3
2015-07-06 23:02:31Learning how to overlay actors using a green screen in #SciPy2015. Look out Hollywood.
2015-07-06 23:03:29Most retractions for #MachineLearning papers due to faulty training/test split algorithms. Proceed with caution #scipy2015
2015-07-06 23:08:21Hi I'm Matt. Ask me about anything! #scipy2015 pic.twitter.com/UWonwsrUBi
2015-07-06 23:09:50cython -a some_cython_code.pyx # generate an annotated html of the code generation. Clues into what might be slow. #scipy2015
2015-07-06 23:30:36.@CjBayesian Plus: you can do this in the @ProjectJupyter notebook Cython cell magic (``%%cython -a``). very efficient workflow #scipy2015
2015-07-06 23:34:37Scientific Computing with Python #scipy2015 - starts with tutorials scipy2015.scipy.org
2015-07-06 23:37:50Hey #scipy2015, let's go to Torchy's Tacos tonight around 6-ish. goo.gl/maps/Uw5FG #swcscipy
2015-07-06 23:38:08scikit-image plus ipython/jupyter notebook widgets == good times at #scipy2015 tutorial pic.twitter.com/k3lStEXm3L
2015-07-06 23:47:19IMHO, @swcarpentry has the best shell tutorial around: swcarpentry.github.io/shell-novice/ #scipy2015 #swcscipy
2015-07-06 23:52:02From Kurt Smith at #scipy2015 Cython tutorial: there’s a new(ish?) tool that does SWIG-like wrapping but uses Cython github.com/xdress/xdress
2015-07-06 23:54:19pre-processing is important and, again, look at your data to figure out the best methods for them! #scipy2015 #MachineLearning
2015-07-06 23:58:33