Update abstractive_model.py

abstractive_model.py

@@ -1,12 +1,22 @@
-from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
+from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, pipeline
 
-# Load the BART tokenizer and model
+# Load the fine-tuned BART tokenizer and model
 tokenizer = AutoTokenizer.from_pretrained("EE21/BART-ToSSimplify")
 model = AutoModelForSeq2SeqLM.from_pretrained("EE21/BART-ToSSimplify")
 
-# Define the abstractive summarization function
-def summarize_with_bart(input_text):
+# Load BART-large-cnn
+pipe = pipeline("summarization", model="facebook/bart-large-cnn")
+
+# Define the abstractive summarization function (fine-tuned BART)
+def summarize_with_bart_ft(input_text):
     inputs = tokenizer.encode("summarize: " + input_text, return_tensors="pt", max_length=1024, truncation=True)
     summary_ids = model.generate(inputs, max_length=200, min_length=50, num_beams=1, early_stopping=False, length_penalty=1)
     summary = tokenizer.decode(summary_ids[0], skip_special_tokens=False)
+    return summary
+
+# Define the abstractive summarization function (BART-large-cnn)
+def summarize_with_bart(input_text):
+    inputs = tokenizer.encode("summarize: " + input_text, return_tensors="pt", max_length=1024, truncation=True)
+    summary_ids = model.generate(inputs, max_length=200, min_length=50, length_penalty=2.0, num_beams=2, early_stopping=True)
+    summary = tokenizer.decode(summary_ids[0], skip_special_tokens=True)
     return summary