|Tolstoyevsky da03dbe866 Fixed eval.dvc and moved its command to a script||1 month ago|
|.dvc||3 months ago|
|docs||3 months ago|
|dvc-example||1 month ago|
|examples||3 months ago|
|fairseq||1 month ago|
|scripts||3 months ago|
|tests||3 months ago|
|.gitignore||4 months ago|
|CONTRIBUTING.md||1 year ago|
|LICENSE||1 year ago|
|PATENTS||1 year ago|
|README.md||2 months ago|
|eval_lm.py||3 months ago|
|fairseq.gif||1 year ago|
|fairseq_logo.png||4 months ago|
|generate.py||1 month ago|
|interactive.py||3 months ago|
|preprocess.py||3 months ago|
|requirements.txt||3 months ago|
|score.py||3 months ago|
|setup.py||3 months ago|
|train.py||1 month ago|
|DVC Managed File|
|DVC Managed File|
Fairseq(-py) is a sequence modeling toolkit that allows researchers and developers to train custom models for translation, summarization, language modeling and other text generation tasks. It provides reference implementations of various sequence-to-sequence models, including:
We also provide pre-trained models for several benchmark translation and language modeling datasets.
Currently fairseq requires PyTorch version >= 1.0.0. Please follow the instructions here: https://github.com/pytorch/pytorch#installation.
If you use Docker make sure to increase the shared memory size either with
--shm-size as command line options to
After PyTorch is installed, you can install fairseq with:
pip install -r requirements.txt python setup.py build develop
The full documentation contains instructions for getting started, training new models and extending fairseq with new model types and tasks.
We provide pre-trained models and pre-processed, binarized test sets for several tasks listed below, as well as example training and evaluation commands.
We also have more detailed READMEs to reproduce results from specific papers:
fairseq(-py) is BSD-licensed. The license applies to the pre-trained models as well. We also provide an additional patent grant.
This is a PyTorch version of fairseq, a sequence-to-sequence learning toolkit from Facebook AI Research. The original authors of this reimplementation are (in no particular order) Sergey Edunov, Myle Ott, and Sam Gross.