This directory contains additional tools and extensions to OpenNMT.

• apply_embeddings.lua applies pre-trained word embeddings to a text file and generates a vector file.
• average_models.lua averages all parameters of multiple models.
• build_vocab.lua generates vocabulary files from text.
• detokenize.lua runs OpenNMT's generic detokenization on a text file.
• embeddings.lua generates pre-trained word embeddings.
• extract_embeddings.lua outputs word embeddings from a model to a text file.
• learn_bpe.lua learns a BPE model.
• release_model.lua releases trained models for inference.
• rest_translation_server.lua runs a simple translation server with a REST API.
• score.lua scores predictions against references with a specific metric.
• tokenize.lua runs OpenNMT's generic tokenization on a text file.
• translation_server.lua runs a simple ZMQ-based translation server.

For more details about these scripts, visit the online documentation or explore available options with the -h flag.