pith. machine review for the scientific record. sign in

arxiv: 1804.00344 · v3 · submitted 2018-04-01 · 💻 cs.CL

Recognition: unknown

Marian: Fast Neural Machine Translation in C++

Marcin Junczys-Dowmunt , Roman Grundkiewicz , Tomasz Dwojak , Hieu Hoang , Kenneth Heafield , Tom Neckermann , Frank Seide , Ulrich Germann
show 4 more authors
Alham Fikri Aji Nikolay Bogoychev Andr\'e F. T. Martins Alexandra Birch
Authors on Pith no claims yet
classification 💻 cs.CL
keywords mariantranslationframeworkmachineneuralachieveautomaticcomputation
0
0 comments X
read the original abstract

We present Marian, an efficient and self-contained Neural Machine Translation framework with an integrated automatic differentiation engine based on dynamic computation graphs. Marian is written entirely in C++. We describe the design of the encoder-decoder framework and demonstrate that a research-friendly toolkit can achieve high training and translation speed.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. SPEC CPU: The Next Generation

    cs.PF 2026-05 unverdicted novelty 7.0

    SPEC CPU 2026 presents a new benchmark suite using open-source apps, expanded multithreading, and Rolling-Round-Robin Rate to address gaps in evaluating heterogeneous multiprogrammed CPU performance.

  2. HuggingFace's Transformers: State-of-the-art Natural Language Processing

    cs.CL 2019-10 accept novelty 6.0

    Hugging Face releases an open-source Python library that supplies a unified API and pretrained weights for major Transformer architectures used in natural language processing.