Shypd.aiThe largest open index of AI toolsGRU4Rec is an open-source Theano implementation of a session-based recommendation algorithm. It utilizes recurrent neural networks and is optimized for GPU execution, processing up to 1500 mini-batches per second.
Trending
About
GRU4Rec is the original Theano implementation of the algorithm described in the "Session-based Recommendations with Recurrent Neural Networks" paper (ICLR 2016) and its follow-up. This open-source tool is specifically optimized for fast execution on GPUs, capable of processing up to 1500 mini-batches per second on a GTX 1080Ti. While official PyTorch and TensorFlow reimplementations exist, this original Theano version is noted for being significantly faster. It provides functionalities for training, evaluating, and saving/loading GRU4Rec models, with detailed configuration options for GPU usage and hyperparameter tuning. The project emphasizes the importance of using this original implementation due to observed flaws and performance issues in third-party versions.
Capabilities
Pricing & Plans
Open Source
Free
FAQs
Trending
