loudinthecloud/pytorch-ntm
A PyTorch implementation of Neural Turing Machines with external memory for learning algorithmic tasks.
Not currently ranked — collecting fresh signals.
star history
This repository provides a vanilla implementation of Neural Turing Machines in PyTorch. An NTM extends traditional sequence processors (like LSTMs) with an external differentiable memory bank, allowing the network to learn algorithmic tasks and maintain internal state longer than standard recurrent networks. The implementation supports batch learning, flexible read/write head configurations, and includes copy and repeat-copy task experiments that reproduce results from the original NTM paper.
Frequently asked
- What is loudinthecloud/pytorch-ntm?
- A PyTorch implementation of Neural Turing Machines with external memory for learning algorithmic tasks.
- Is pytorch-ntm open source?
- Yes — loudinthecloud/pytorch-ntm is open source, released under the BSD-3-Clause license.
- What language is pytorch-ntm written in?
- loudinthecloud/pytorch-ntm is primarily written in Jupyter Notebook.
- How popular is pytorch-ntm?
- loudinthecloud/pytorch-ntm has 610 stars on GitHub.
- Where can I find pytorch-ntm?
- loudinthecloud/pytorch-ntm is on GitHub at https://github.com/loudinthecloud/pytorch-ntm.