lyken17 / ignite Goto Github PK

Ignite is a high-level library to help with training neural networks in PyTorch.

Note: Ignite is currently in alpha, and the API is subject to change. A first release coming soon, where we will stabilize the API. Stay tuned.

From source:

python setup.py install

Anaconda & pip coming soon

API documentation and an overview of the library coming soon.

Please check out the examples to see how to use ignite to train various types of networks, as well as how to use visdom or tensorboardX for training visualizations.

We appreciate all contributions. If you are planning to contribute back bug-fixes, please do so without any further discussion. If you plan to contribute new features, utility functions or extensions, please first open an issue and discuss the feature with us.

Please see the contribution guidelines for more information.

As always, PRs are welcome :)

Ignite, in spirit is very similar to torchnet (and was inspired by torchnet).

The main differences with torchnet is the level of abstraction for the user. Ignite's higher level of abstraction assumes less about the type of network (or networks) that you are training, and we require the user to define the closure to be run in the training and validation loop. In contrast to this, torchnet creates this closure internally based on the network and optimizer you pass to it. This higher level of abstraction allows for a great deal more of flexibility, such as co-training multiple models (i.e. GANs) and computing/tracking multiple losses and metrics in your training loop.

Ignite also allows for multiple handlers to be attached to events, and a finer granularity of events in the engine loop.