Added application example

README.md

@@ -55,3 +55,30 @@ After multiple test runs of finetuning the present model was further trained usi
 - weight_decay: 0.1
 - metric_for_best_model: precision
 
+### Example: Direct model evaluation
+
+```python
+
+from transformers import (
+    AutoModelForSequenceClassification,
+    AutoTokenizer,
+    pipeline,
+)
+
+comment = "The preprocessed comment to classify"
+tokenizer = AutoTokenizer.from_pretrained(model_path)
+tokenizer.pad_token = tokenizer.eos_token
+model = AutoModelForSequenceClassification.from_pretrained(model_path)
+pipe = pipeline("text-classification", model=model, tokenizer=tokenizer)
+
+result = pipe(comment)
+label = result[0]["label"]
+if label == "LABEL_1":
+    has_mention = True
+elif label == "LABEL_0":
+    has_mention = False
+
+print(f"Comment includes mention {has_mention}")
+```
+
+