Update app.py

app.py

@@ -1,23 +1,29 @@
-# Import necessary libraries
-import blurr
-import gradio as gr
+import transformers
+from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
+from gradio import Interface as gr
 from pathlib import Path
 from fastai.text.all import *
-from blurr.text.data.all import *
-from blurr.text.modeling.all import *
-
-# Manually download and prepare SQuAD dataset
 from datasets import load_dataset
+
+# Download and prepare SQuAD dataset (not used directly here)
 squad = load_dataset("squad")
 
-# Load the learner without using SQuAD
-inf_learn = load_learner(fname=Path("laptop_summarizer_1.pkl"), trust_remote_code=True)
+# Load the pre-trained summarization model (adjust model name as needed)
+model_name = "facebook/bart-base"  # Choose a suitable summarization model
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
 
-# Define a function to generate summaries using your model
+# Define a function to generate summaries using the model
 def generate_summary(input_text):
-    prediction = inf_learn.blurr_generate(input_text)
-    generated_text = prediction[0]['generated_texts']
-    return generated_text
+  # Tokenize the input text
+  inputs = tokenizer(input_text, return_tensors="pt")
+
+  # Generate summary using the pre-trained model
+  output = model.generate(**inputs)
+
+  # Decode the generated tokens back to text
+  summary_text = tokenizer.batch_decode(output, skip_special_tokens=True)[0]
+  return summary_text
 
 # Create an interface for the model
 interface = gr.Interface(
@@ -30,4 +36,4 @@ interface = gr.Interface(
 )
 
 # Start the Gradio app
-interface.launch(inline=True, trust_remote_code=True)
+interface.launch(inline=True)