How AI describes CTranslate2

CTranslate2 is a fast offline neural machine translation inference engine that supports text translation, generation, encoding, and speech recognition. The documentation provides installation, usage guides, and API references, including tools for model conversion, quantization, decoding, and tips on memory management and performance. It supports multiple frameworks and runtimes—Fairseq, Marian, OpenNMT-py, OpenNMT-tf, OPUS-MT, Transformers—with Python APIs, plus hardware guidance and versioning information.