diff options
| -rw-r--r-- | readme.md | 2 |
1 files changed, 1 insertions, 1 deletions
@@ -54,7 +54,7 @@ We include a pre-trained character-level tokenizer for the german language in th If the path to the tokenizer you specified in the `config.yaml` file does not exist or is None (~), a new tokenizer will be trained on the training data. -## Decoder : +## Decoder There are two options for the decoder: - greedy - beam search with language model |