Skip to content

Update/speed up to the lstm code #71

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 7 commits into from
Feb 3, 2015
Merged

Update/speed up to the lstm code #71

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
9018392
pep8
nouiz Jan 14, 2015
906a30e
Small comment update
nouiz Jan 16, 2015
26914e4
Add a parameter to make the test set smaller
nouiz Jan 16, 2015
ec112fa
use floatX for lstm
nouiz Jan 21, 2015
6a32f03
sort dataset by length to speed up error computation
nouiz Jan 31, 2015
2f14fbf
Try to repair travis failure to install our dependency
nouiz Jan 31, 2015
6e38783
Keep random subset of the test set
nouiz Feb 3, 2015
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Next Next commit
Small comment update
  • Loading branch information
@nouiz
nouiz committed Jan 31, 2015
commit 906a30e37df0801a5d3c847affebe46453075ae7
1 change: 1 addition & 0 deletions code/imdb.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,7 @@ def prepare_data(seqs, labels, maxlen=None):
if maxlen is set, we will cut all sequence to this maximum
lenght.

This swap the axis!
"""
# x: a list of sentences
lengths = [len(s) for s in seqs]
Expand Down
10 changes: 4 additions & 6 deletions code/lstm.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -481,12 +481,10 @@ def train_lstm(
y = [train[1][t] for t in train_index]
x = [train[0][t]for t in train_index]

# Get the data in numpy.ndarray formet.
# It return something of the shape (minibatch maxlen, n samples)
x, mask, y = prepare_data(x, y, maxlen=maxlen)
if x is None:
print 'Minibatch with zero sample under length ', maxlen
continue
# Get the data in numpy.ndarray format
# This swap the axis!
# Return something of shape (minibatch maxlen, n samples)
x, mask, y = prepare_data(x, y)
n_samples += x.shape[1]

Copy link
Member

@lamblin lamblin Feb 2, 2015

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why remove this check? prepare_data can still return (None, None, None).

Copy link
Member Author

@nouiz nouiz Feb 3, 2015

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

as I do not pass anymore the maxlen parameter, prepare_data won't return None.

Copy link
Member

@lamblin lamblin Feb 3, 2015

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Oh, OK.

cost = f_grad_shared(x, mask, y)
Expand Down