[nlp]
lang = "eds"
pipeline = ["ner"]

[nlp.tokenizer]
@tokenizers = "eds.tokenizer"

[components]

[components.ner]
@factory = "eds.ner_crf"
target_span_getter = "gold_spans"
labels = ["anatomie", "date", "dose", "duree", "examen", "frequence", "mode", "moment", "pathologie", "sosy", "substance", "traitement", "valeur"]
infer_span_setter = true
mode = "joint"
window = 40
stride = 20

[components.ner.embedding]
@factory = "eds.text_cnn"
kernel_sizes = [3]

[components.ner.embedding.embedding]
@factory = "eds.transformer"
model = "./ner/embedding/embedding"
window = 128
stride = 96

[components.ner.span_setter]
ents = true
* = true
gold_spans = ["anatomie", "date", "dose", "duree", "examen", "frequence", "mode", "moment", "pathologie", "sosy", "substance", "traitement", "valeur"]