Update README.md

README.md

@@ -8,6 +8,16 @@ datasets:
 co2_eq_emissions: 1.06099358268878
 ---
 
+# Prediction of sentence "nature" in a French political sentence
+
+This model aims at predicting the nature of a sentence in a French political sentence.
+The predictions fall in three categories:
+- `problem`: the sentence describes a problem (usually to be tackled by the speaker), for example _il y a dans ce pays une fracture_ (J. Chirac)
+- `solution`: the sentences describes a solution (typically part of a political programme), for example: _on va donc nettoyer, au propre comme au figuré, la cité des 4000_ (N. Sarkozy)
+- `other`: the sentence does not belong to any of these categories, for example: _vive la République, vive la France_
+
+This model was trained using AutoNLP based on sentences extracted from a mix of political tweets and speeches.
+
 # Model Trained Using AutoNLP
 
 - Problem type: Multi-class Classification
@@ -49,4 +59,8 @@ tokenizer = AutoTokenizer.from_pretrained("mazancourt/politics-sentence-classifi
 inputs = tokenizer("Il y a dans ce pays une fracture", return_tensors="pt")
 
 outputs = model(**inputs)
+
+# Category can be "problem", "solution" or "other"
+category = outputs[0]["label"]
+score = outputs[0]["score"]
 ```