test_that("Can initialize a tokenizer different models", { models <- list( model_bpe$new(), model_wordpiece$new(), model_unigram$new(), model_unigram$new(vocab = c("hello" = 10, "bye" = 12), 1L), model_unigram$new(vocab = list("hello" = 10, "bye" = 12), 1L) ) for (model in models) { tok <- tokenizer$new(model) expect_true(inherits(tok, "tok_tokenizer")) } })