Upload README.md with huggingface_hub

README.md

@@ -33,12 +33,7 @@ The following table compares **THIVLVC** against major industry standards across
 
 ## Usage
 
-Installation:
-```bash
-pip install transformers torch
-```
-
-Basic usage in Python:
+**Important**: For best results, especially on short sentences or fragments, use **beam search** (`num_beams=5`).
 
 ```python
 from transformers import AutoTokenizer, T5ForConditionalGeneration
@@ -49,11 +44,13 @@ model = T5ForConditionalGeneration.from_pretrained(model_name)
 
 def lemmatize(text):
     inputs = tokenizer(text, return_tensors="pt")
-    outputs = model.generate(**inputs, max_length=128)
+    # Using beam search (num_beams=5) for better accuracy
+    outputs = model.generate(**inputs, max_length=128, num_beams=5, early_stopping=True)
     return tokenizer.decode(outputs[0], skip_special_tokens=True)
 
 # Example
-print(lemmatize("Amorem canat"))
+print(lemmatize("Amorem canat")) 
+# Expected Output: "amor cano"
 ```
 
 ## Dataset and Training