Lasagne: First release.

By Sander Dieleman, Michael Heilman, J. Kelly, M. Thoma, K. Rasul, Eric Battenberg, Hendrik J. Weideman, Søren Kaae Sønderby, instagibbs, Britefury and 13 more. 2015

Subfields: ML Systems & Hardware

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In ML Systems & Hardware

Citations in the corpus

The One Hundred Layers Tiramisu: Fully Convolutional DenseNets for Semantic Segmentation (2016, background): “Vid and Gatech), without neither additional post-processing, pretraining, nor including temporal information. Aknowledgements The authors would like to thank the developers of Theano [32] and Lasagne [7]. Special thanks to Fred´ ´eric Bastien for his work assessing the compilation issues. Thanks to ”
Quantized Neural Networks: Training Neural Networks with Low Precision Weights and Activations (2016, background): “, 2013a) and Lasagne (Dieleman et al., 2015), two deep learning libraries built on the top of Theano.”
Theano: A Python framework for fast computation of mathematical expressions (2016, background): “For instance, machine learning and deep learning packages, such as Pylearn2 [8], Blocks [9], Lasagne [10], and Keras [11], have been developed with the goal of making it easier to express the architecture of deep learning models, and training algorithms, as mathematical expressions to be evaluated…”
Auxiliary Deep Generative Models (2016, methodology): “The models are implemented in Python using Theano (Bastien et al., 2012), Lasagne (Dieleman et al., 2015) and Parmesan libraries 1 .”
Temporal Ensembling for Semi-Supervised Learning (2016, methodology): “Implementation Our implementation is written in Python using Theano (Theano Development Team, 2016) and Lasagne (Dieleman et al., 2015), and is available at https://github.com/smlaine2/tempens.”
Neural Photo Editing with Introspective Adversarial Networks (2016, methodology): “Our models are implemented in Theano (Team, 2016) with Lasagne (Dieleman et al., 2015).”
BinaryConnect: Training Deep Neural Networks with binary weights during propagations (2015, methodology): “We also thank the developers of Pylearn2 [44] and Lasagne [45], two Deep Learning libraries built on the top of Theano.”
Ladder Variational Autoencoders (2016, methodology): “The models were implemented using the Theano [19], Lasagne [4] and Parmesan 5 frameworks.”
Dynamic Filter Networks (2016, methodology): “We use Theano [2] based library Lasagne [3] to implement all the experiments.”
Exploiting Unintended Feature Leakage in Collaborative Learning (2018, methodology): “ental setup. Our experiments were performed on a workstation running Ubuntu Server 16.04 LTS equipped with a 3.4GHz CPU i7-6800K, 32GB RAM, and an NVIDIA TitanX GPU card. We use MxNet [8] and Lasagne [12] to implement deep neural networks and Scikit-learn [41] for conventional ML models. The source ”
Machine Learning Models that Remember Too Much (2017, methodology): “training algorithm has a hard-coded secret that can be used as the key for a pseudorandom function or encryption. All ML models and attacks were implemented in Python 2.7 with Theano [70] and Lasagne [20]. The experiments were conducted on a machine with two 8-core Intel i7-5960X CPUs, 64GB RAM, and”
Early Methods for Detecting Adversarial Images (2016, methodology): “2 We demonstrate the results in Figure 4 using a pretrained VGG-16 model (Dieleman et al., 2015; Simonyan & Zisserman, 2015).”
Learning how to explain neural networks: PatternNet and PatternAttribution (2017, methodology): “We used Theano (Bergstra et al., 2010) and Lasagne (Dieleman et al., 2015) for our implementation.”
Automated Machine Learning: State-of-The-Art and Open Challenges (2019, methodology): “Auto-net accesses deep learning techniques from Lasagne Python deep learning library [29].”
Towards a Neural Statistician (2016, methodology): “We primarily use the Theano [29] framework with the Lasagne [4] library, but the ﬁnal experiments with face data were done using Tensorﬂow [1].”
Phased LSTM: Accelerating Recurrent Network Training for Long or Event-based Sequences (2016, methodology): “ich were then fully-connected to 256 neurons, and ﬁnally fullyconnected to the 10 output classes. The event pixel address was used to produce a 40-dimensional embedding via a learned embedding matrix [9], and combined with the polarity to produce the input. Therefore, the network architecture was 41”
Online Structured Laplace Approximations For Overcoming Catastrophic Forgetting (2018, methodology): “re over a diagonal one. We also run EWC with a Kronecker factored approximation, even though the original method is based on a diagonal one. We implement our experiments using Theano [39] and Lasagne [3] software libraries. 4.1 Permuted MNIST As a ﬁrst experiment, we test on a sequence of permutatio”