| --- |
| { |
| "language": "fr", |
| "license": "apache-2.0", |
| "tags": [ |
| "medical", |
| "icd10", |
| "document-classification", |
| "text-classification" |
| ], |
| "pipeline_tag": "text-classification" |
| } |
| --- |
| |
| # DP classifier (CIM-10, fr) |
|
|
| Modèle de classification du Diagnostic Principal (40 codes CIM-10) fine-tuné sur des comptes rendus fictifs. |
|
|
| ## Utilisation (Transformers) |
| ```python |
| from transformers import AutoTokenizer, AutoModelForSequenceClassification |
| |
| repo = "clairedhx/camembert_dp_ft_top40_bdx" |
| tok = AutoTokenizer.from_pretrained(repo, use_fast=True) |
| model = AutoModelForSequenceClassification.from_pretrained(repo) |
| |
| x = tok("Douleur thoracique avec sus-décalage ST...", return_tensors="pt") |
| pred_id = int(model(**x).logits.argmax(-1)) |
| print(model.config.id2label[pred_id]) |
| ``` |
|
|
|
|