www-announce mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Wang Wei <wang...@apache.org>
Subject [ANNOUNCE] Apache SINGA (incubating) 0.2.0 release
Date Fri, 15 Jan 2016 03:42:37 GMT

We are pleased to announce that Apache SINGA (incubating) 0.2.0 is released.

SINGA is a general distributed deep learning platform for training big deep
learning models over large datasets. It is designed with an intuitive
programming model based on the layer abstraction. SINGA supports a wide
variety of popular deep learning models.

The release is available at:

The main features of this release include

* Training on GPU  -- enabling training of complex models on a single node
with multiple GPU cards.
* Hybrid neural net partitioning -- supporting data and model parallelism
at the same time.
* Python wrapper -- making it easier to configure jobs, including neural
net and SGD algorithm.
* RNN model and BPTT algorithm -- supporting applications based on RNN
models, e.g., GRU.
* Cloud software integration, including Mesos, Docker and HDFS.
* Visualization of neural net structure and layer information -- helpful
for debugging.
* Linear algebra functions and random functions against Blobs and raw data
* New layers, including SoftmaxLayer, ArgSortLayer, DummyLayer, RNN layers
and cuDNN layers.
* Update Layer class -- for carrying multiple data/grad Blobs.
* Extract features and test performance for new data by loading previously
trained model parameters.
* Add Store class for IO operations

We look forward to hearing your feedbacks, suggestions, and contributions
to the project (http://singa.apache.org/develop/schedule.html).

On behalf of the SINGA team,
Wei Wang

View raw message