Olivier Grisel (ogrisel)
- Website
- http://oliviergrisel.name/
- Location
- Paris
- Member since
Recent activity
-
Olivier Grisel
-
commented on pull request #2 in pyleargist
-
began following Nuxeo FUSE
-
committed to libsgd
8a4d719ee2a7fix indent1a55de402d15temporary hack to work with wikiformat pictures -
deleted mdp
-
deleted mdp
-
began following pimpy
-
committed to gumby
13f5c9ac334dexperimental branch to use more numpy arrays instead of python list to spare memory and calls to malloc262f72cf4e56fix style + train a model by default91984b1e4644smaller default dimensions + added timing information9b72a63365c8ignore ctags index filed6a3810f65ecuse 32bit floats by default to spare some memory898a220f76c2make it possible to store the training set of a 1 layer experiment for later usagef5d27b85dfe4make Layer.build_training_set() public -
began following stellaris
-
Repositories
-
cloudmole
Text-mining on the cloud with machine learning and Hadoop MapReduce in Scala
- Privacy
- Public repository
- Followers
-
ctypes_sse
Minimal sample program to reproduce a segmentation fault when calling a function
with SSE intrinsics from python using ctypes.
- Privacy
- Public repository
- Followers
-
gumby
Experimental project to implement a cortical inference model with python / scipy and support vector machines. The goal is too devise a new time series forecasting ML tool able to infer abstract representations of the input data to build better estimates.
The main idea is to build a layered structure of simple predictive units (linear autoregressive models, logistic regression units, SVM with kernels or two-layers perceptron locally trained with backprop) and make the prediction errors flow up from the lowest level units while the upper level units give their estimates of future prediction error to the lowest units.
The project is still in an early alpha development stage. Please feel free to take a look at the source code and contact us by IRC on #gumby on freenode.org.
- Privacy
- Public repository
- Followers
-
libsgd
libsgd is a SSE optimized C library with Python bindings (using ctypes)
The goal of libsgd is to provide an optimized implementation of
backpropagatable modules as a base for efficient machine learning
algorithms such as feed forward multi layer perceptron, stacked
autoencoders, linear support vector machines solved in the primal and
potentially more.
- Privacy
- Public repository
- Followers
-
og.lab
Personnal branch to host experimental code not yet part of any identified project per se.
- Privacy
- Public repository
- Followers
-
pyleargist
GIST Image descriptor for scene recognition.
cython / numpy wrapper of the PSL'd C implementation by the LEAR team:
http://lear.inrialpes.fr/software
- Privacy
- Public repository
- Followers
-
pyprof2calltree
Script to help visualize profiling data collected with the cProfile python module with the kcachegrind graphical calltree analyser.
This is a rebranding of the venerable http://www.gnome.org/~johan/lsprofcalltree.py script to make it easier to distribute on pypi and behave more like the scripts of the debian kcachegrind-converters package.
The final goal is to make it part of the official upstream kdesdk.
- Privacy
- Public repository
- Followers
-
virtualkeyring
- Privacy
- Public repository
- Followers