Search Results for "translation memory"

...No such runtime translation exists in Micronaut Data and this work is carried out by the Micronaut compiler at compilation time.

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. We provide reference implementations of various sequence modeling papers. Recent work by Microsoft and Google has shown that data parallel training can be made significantly more efficient by sharding the model parameters and optimizer state across data parallel workers. These ideas are encapsulated in the...

Concordia - Roman goddess of agreement. Concordance searcher - tool for translators who need their translations to "agree" with one standard. Concordia is a C++ library for fast text lookup in large corpora. It uses a RAM stored index, which takes up approximately 600MB of memory for a corpus of 2 million sentences. It is based on the idea of a suffix array, enhanced by the presence of other auxiliary data structures. The effects are stunning - Concordia is able to do simple substring...

...It parses C, maps C types and constructs into Rust equivalents, and generates code that compiles under rustc, introducing unsafe only when necessary. The tool seeks to produce readable Rust that a developer can then refine by hand, rather than a perfect one-to-one mechanical translation. It handles common C features such as pointers, structs, enums, arrays, and function calls, while flagging areas that need attention during the migration. Preprocessor handling and tricky macro patterns are approached pragmatically, aiming for working output over exhaustive transformation. As a proof-of-concept, it demonstrates how automated tooling can accelerate moving from legacy C to a memory-safe language without a full rewrite.