Update app.py

app.py

@@ -1,23 +1,79 @@
+
 import gradio as gr
 from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
+import nltk
+import torch
+from textblob import TextBlob
 
-model_name = "ramsrigouthamg/t5_paraphraser"
-tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast=False)
-model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
+nltk.download("punkt")
 
-def paraphrase_text(text):
-    input_text = "paraphrase: " + text + " </s>"
-    encoding = tokenizer.encode_plus(
-        input_text, padding="max_length", return_tensors="pt", max_length=256, truncation=True
-    )
-    outputs = model.generate(
-        input_ids=encoding["input_ids"],
-        attention_mask=encoding["attention_mask"],
+model = AutoModelForSeq2SeqLM.from_pretrained("ramsrigouthamg/t5_paraphraser")
+tokenizer = AutoTokenizer.from_pretrained("ramsrigouthamg/t5_paraphraser", use_fast=False)
+
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+model = model.to(device)
+
+def split_into_sentences(text):
+    return nltk.tokenize.sent_tokenize(text)
+
+def paraphrase_sentence(sentence, creativity):
+    text = "paraphrase: " + sentence + " </s>"
+    encoding = tokenizer.encode_plus(text, padding="max_length", return_tensors="pt", max_length=256, truncation=True)
+    input_ids, attention_mask = encoding["input_ids"].to(device), encoding["attention_mask"].to(device)
+
+    output = model.generate(
+        input_ids=input_ids,
+        attention_mask=attention_mask,
         max_length=256,
-        num_return_sequences=1,
-        temperature=1.5,
+        do_sample=True,
+        top_k=120,
+        top_p=creativity,
+        early_stopping=True,
+        num_return_sequences=1
     )
-    return tokenizer.decode(outputs[0], skip_special_tokens=True)
+    paraphrased = tokenizer.decode(output[0], skip_special_tokens=True, clean_up_tokenization_spaces=True)
+    return paraphrased
+
+def correct_grammar(text):
+    blob = TextBlob(text)
+    return str(blob.correct())
+
+def paraphrase_text(text, creativity=0.9, tone="neutral", improve_grammar=True, batch_size=5):
+    sentences = split_into_sentences(text)
+    results = []
+
+    for sentence in sentences:
+        para = paraphrase_sentence(sentence, creativity)
+        if improve_grammar:
+            para = correct_grammar(para)
+        results.append(para)
 
-demo = gr.Interface(fn=paraphrase_text, inputs="text", outputs="text", title="Free Paraphraser")
-demo.launch()
+    return " ".join(results)
+
+with gr.Blocks() as demo:
+    gr.Markdown("<h1>Paraphrasing Tool</h1><p>AI-powered rewriting with grammar improvement and creativity controls.</p>")
+
+    with gr.Row():
+        input_text = gr.Textbox(lines=8, label="Enter Text")
+        output_text = gr.Textbox(lines=8, label="Paraphrased Output")
+
+    creativity = gr.Slider(0.1, 1.0, value=0.9, step=0.05, label="Creativity (top_p)")
+    improve_grammar = gr.Checkbox(value=True, label="Improve Grammar")
+    tone = gr.Radio(["neutral", "formal", "casual"], label="Tone (placeholder)", value="neutral")
+    batch_size = gr.Slider(1, 10, value=5, step=1, label="Batch Size")
+
+    run_button = gr.Button("Paraphrase")
+
+    run_button.click(
+        paraphrase_text,
+        inputs=[input_text, creativity, tone, improve_grammar, batch_size],
+        outputs=output_text
+    )
+
+if __name__ == "__main__":
+    demo.launch(
+        server_name="0.0.0.0",
+        server_port=7860,
+        show_api=True,
+        favicon_path="favicon.ico"
+    )