# English model
languages:  [
  "en"
]

# Model has 512 positional embeddings, but only trained on sequences up to 256 tokens
max_length: 256

# Not actually used, but included for pipeline generality
pre_labels: [
  "<NULL>",
  "¿",
]

post_labels: [
  "<NULL>",
  "<ACRONYM>",
  ".",
  ",",
  "?",
]