Update app.py

app.py

@@ -1,25 +1,63 @@
-import streamlit as st
 import torch
-from transformers import pipeline, AutoTokenizer, AutoModelForSequenceClassification
+import streamlit as st
+from transformers import pipeline, AutoTokenizer, AutoModelForCausalLM
 
-# Load pre-trained model and tokenizer
-model_name = "distilbert-base-uncased"
+model_name = "gpt2"
 tokenizer = AutoTokenizer.from_pretrained(model_name)
-model = AutoModelForSequenceClassification.from_pretrained(model_name)
-
-# Create Streamlit app
-st.title("Hugging Face Transformers + Streamlit Example")
-
-# Define a function to use the model for prediction
-@st.cache(allow_output_mutation=True)
-def run_model(input_text):
-    inputs = tokenizer(input_text, return_tensors="pt")
-    outputs = model(**inputs)
-    predictions = torch.softmax(outputs.logits, dim=1)
-    return predictions
-
-# Streamlit interface
-user_input = st.text_input("Enter text:", "Type Here...")
-if st.button("Predict"):
-    predictions = run_model(user_input)
-    st.write(f"Predictions: {predictions}")
+model = AutoModelForCausalLM.from_pretrained(model_name)
+
+generator = pipeline(
+    "text-generation",
+    model=model,
+    tokenizer=tokenizer,
+    device=0 if torch.cuda.is_available() else -1
+)
+
+# Function for text generation with filtering of repeated sequences
+def generate_text(prompt, section, max_length=200, temperature=0.7, top_k=50, repetition_penalty=1.2):
+    return generator(
+        f"{section} - {prompt}",
+        max_length=max_length,
+        do_sample=True,
+        top_k=top_k,
+        temperature=temperature,
+        repetition_penalty=repetition_penalty,
+        num_return_sequences=1,
+        eos_token_id=tokenizer.eos_token_id,
+    )[0]["generated_text"]
+
+# Streamlit app
+st.title("AI-Generated Blog Post")
+
+# Keyword selection input
+keywords_input = st.text_input("Step 1: Keyword Selection (Separate keywords with commas)","Artificial Intelligence")
+keywords = [word.strip() for word in keywords_input.split(',')]
+
+# Display generated content on button click
+if st.button('Generate Article'):
+    if keywords_input:
+
+        generated_text = " ".join(keywords)
+        intro_text = generate_text(generated_text, "Introduction", max_length=200, temperature=0.7, top_k=50)
+        body_text = generate_text(generated_text, "Body", max_length=500, temperature=0.7, top_k=50)
+        conclusion_text = generate_text(generated_text, "Conclusion", max_length=150, temperature=0.7, top_k=50)
+
+        # Displaying the sections with adjusted parameters
+        st.header("Introduction")
+        st.write(intro_text)
+
+        st.header("Body")
+        st.write(body_text)
+
+        st.header("Conclusion")
+        st.write(conclusion_text)
+    else:
+        st.warning("Please input keywords to generate content.")
+
+# Sidebar with instructions
+st.sidebar.title("Instructions")
+st.sidebar.write(
+    "1. Enter keywords related to the topic you want to generate content about."
+    "\n2. Click 'Generate Article' to create the AI-generated blog post."
+    "\n3. Explore the Introduction, Body, and Conclusion sections of the generated content."
+)