dnouri / nolearn

Combines the ease of use of scikit-learn with the power of Theano/Lasagne

http://danielnouri.org/notes/2014/12/17/using-convolutional-neural-nets-to-detect-facial-keypoints-tutorial/

dnouri/nolearn Issues

nolearn now on conda-forge
Updated 3 years ago1
python 3 support not working with Lasagne?
Closed 5 years ago12
TypeError: Failed to instantiate <class 'lasagne.layers.pool.MaxPool2DLayer'> with args {'name': 'pool1', 'ds': (2, 2), 'incoming': <lasagne.layers.conv.Conv2DLayer object at 0x7ff765fa29e8>}. Maybe parameter names have changed?
Updated 5 years ago
enable to reproduce loss value of training when predicting CNN
Updated 6 years ago1
Enable to reproduce the last value of trainning when predicting CNN
Closed 6 years ago
Dependency on both backends in requirements.txt switches off GPU support
Updated 6 years ago3
OSError: could not read bytes when trying to fetch mldata
Closed 8 years ago2
Install nolearn with Lasagne dependance not working
Updated 7 years ago2
Future issue with sklearn.cross_validation
Updated 7 years ago
Weights sum up to zero
Updated 7 years ago
Does the partial_fit method reset the parameters?
Closed 8 years ago7
'NeuralNet' object has no attribute 'layers_'
Closed 8 years ago1
Bug when using Lasagne `mask_input` parameter
Closed 8 years ago
nolearn is not installing
Updated 8 years ago
remove('trainable') Lasagne's command doesn't work in nolearn
Updated 8 years ago6
Bug in calculating average scores
Updated 8 years ago
Trained on GPU, inference on CPU doesn't make sense
Closed 8 years ago
flip_filters and pad parameter not used by NeuralNet's class
Updated 8 years ago5
CUDA error, possibly related to network size?
Closed 8 years ago2
A replayable fit() method - diff/patch attached
Updated 8 years ago1
Sklearn GridSearchCV compatible assigning weights for unbalanced classes
Closed 8 years ago3
Remove explicit `objective_loss_function` from `NeuralNet`
Closed 8 years ago2
'RuntimeError: maximum recursion depth exceeded' when trying to serialize model
Closed 8 years ago6
RememberBestWeights does not honor the verbose parameter
Closed 8 years ago2
TrainSplit(eval_size=0.0) and BatchIterator(shuffle=True) crashes
Updated 8 years ago1
2 networks with shared layers
Closed 8 years ago5
TypeError when l1 regularization in objective: unsupported operand type(s) for *: 'TensorVariable' and 'function'
Closed 8 years ago3
Feature request: default score for NeuralNet regression matching sklearn
Closed 8 years ago
AttributeError: 'NeuralNet' object has no attribute 'scores_train'
Closed 8 years ago2
Multiple Input layers?
Closed 8 years ago1
Convolutional Autoencoder NaN
Closed 8 years ago2
Citing nolearn
Closed 8 years ago3
Best way to reuse layers from another model for a new model
Closed 8 years ago2
scores_train/valid take different parameters than custom_scores
Closed 8 years ago1
Include optional colormaps in "nolearn/lasagne/visualize.py" ?
Updated 8 years ago3
using lasagne.layers.BatchNormLayer
Closed 8 years ago1
Prediction using learned weights
Closed 8 years ago2
Getting Type Error while training Convnet using nolearn.lasagne
Closed 8 years ago7
Get output/ score of dense_layer
Closed 8 years ago2
Train class probabilities instead of hard labels
Closed 8 years ago2
Having problems loading parameters
Closed 8 years ago3
score broken if batch_iterator_test doesn't yield same order / size as y
Updated 8 years ago
Allow overriding `accuracy` metric
Updated 8 years ago1
override PrintLog
Closed 8 years ago1
Is there an option in NeuralNet to change the training data at each epoch ?
Closed 8 years ago4
Dealing with large datasets:
Closed 8 years ago
Missing parentheses in call to 'print'
Closed 8 years ago1
Network not learning
Closed 8 years ago3
Large batch_size leads to inaccurate validation loss calculation
Closed 8 years ago2
PrintLayerInfo handler doesn't print layer indices correctly
Closed 8 years ago