fuel
fuel copied to clipboard
mila-iqia
A data pipeline framework for machine learning
These are out of core array and table abstractions for a variety of backends (including numpy). Also have chunking and streaming capability. See some of these : blaze/dask#138 http://blaze.pydata.org/ Just...
Using the pip install git command, it fails with the following output: ``` running install running bdist_egg running egg_info creating fuel.egg-info writing fuel.egg-info\PKG-INFO writing requirements to fuel.egg-info\requires.txt writing entry points...
When creating your own extra_converters, it appears it is necessary to add them to the `__init__.py` in the `extra_converters` module directory, but this isn't in the tutorial documentation on `extra_converters`....
One of the goals for factoring Fuel out of Blocks was to be able to re-use it as a new dataset back-end in Pylearn2. Here is a proposal and a...
I noticed this while trying to figure out how to correctly preserve axis labels in transformer streams. Basically, the current behaviour is wrong with respect to `DataStream`, since if the...
Consider this code: ``` python In [1]: from fuel.datasets import MNIST In [2]: import numpy In [3]: first_five_targets = MNIST('train').get_data(request=range(5))[-1] # They happen to be 5, 0, 4, 1, 9...
It is often possible to determine from the iteration scheme and the dataset how many batches will be returned (e.g. `ceil(num_examples / batch_size)`). This information could be useful (e.g. for...
It seems to happen reliably on one machine with Ubuntu 32-bit, where hdf5 is installed by Anaconda. I tried all the available versions of hdf5/pytables/h5py available on Anaconda, all exhibit...
For example, for `SourcewiseTransformer`.
