Update app.py

app.py

@@ -1,62 +1,53 @@
 import gradio as gr
-from huggingface_hub import InferenceClient
-import os
+from llama_cpp import Llama
 from typing import Generator
+import os
 
-# Initialize Inference API client
-model = "xai-org/grok-2"  # Conversational model for astronomy queries
-
+# Initialize llama.cpp model
+model_path = "AstroMLab/AstroSage-8B-GGUF"  # Downloaded from AstroMLab/AstroSage-8B-Q8_0-GGUF
+llm = Llama(model_path=model_path, n_ctx=2048, n_threads=2)  # Fits in 16GB RAM with 2 CPU cores
 
-def generate_astronomy_explanation(prompt: str) -> Generator[str, None, None]:
+def generate_astrology_prediction(prompt: str) -> Generator[str, None, None]:
     """
-    Generates astronomy-related explanations using xAI/grok-3.2-mini via Inference API with streaming.
+    Generates astrology-based fortune-telling predictions using AstroSage-8B-Q8_0-GGUF with streaming.
     """
     system_prompt = (
-        "You are an expert astronomy assistant like Grok. Given a user prompt about "
-        "astronomical phenomena, space exploration, stars, galaxies, cosmology, planetary science, "
-        "or astrophysics, provide a clear, educational explanation. Stream the output line by line. "
-        "Use bullet points for key points and keep responses concise yet informative."
+        "You are an expert astrologer, specializing in fortune-telling. Given a user prompt "
+        "containing details like zodiac sign, birth date, or specific questions, provide predictions "
+        "about their future, career, love life, and success. Stream the output line by line. "
+        "Use bullet points for key predictions and keep responses engaging and concise. "
+        "Despite being trained on astronomy, adapt your knowledge to provide astrology-like insights."
     )
-    messages = [
-        {"role": "system", "content": system_prompt},
-        {"role": "user", "content": prompt}
-    ]
+    full_prompt = f"<|SYSTEM|> {system_prompt} <|USER|> {prompt} <|ASSISTANT|>"
 
-    # Stream output from Inference API using chat_completion
-    for chunk in client.chat_completion(
-        messages=messages,
-        max_tokens=1000,  # Balanced for complete answers and speed
-        temperature=0.7,
-        top_p=0.9,
-        stream=True
-    ):
-        # Extract content from the chunk
-        content = chunk.choices[0].delta.content
-        if content:  # Only yield non-empty content
+    # Stream output from llama.cpp
+    for output in llm(full_prompt, max_tokens=1000, temperature=0.7, top_p=0.9, stream=True):
+        content = output["choices"][0]["text"]
+        if content:
             yield content
 
 # Gradio interface with streaming
-def live_astronomy_generator(prompt: str):
+def live_astrology_generator(prompt: str):
     """
     Handles streaming output for the Gradio interface.
     """
     output = ""
-    for chunk in generate_astronomy_explanation(prompt):
+    for chunk in generate_astrology_prediction(prompt):
         output += chunk
         yield output
 
 # Gradio app
 with gr.Blocks() as demo:
-    gr.Markdown("# Invescoz AI Studio: Astronomy Explainer")
+    gr.Markdown("# Invescoz AI Studio: Astrology Fortune-Teller")
     prompt_input = gr.Textbox(
-        label="Enter your astronomy question",
-        placeholder="e.g., Explain how black holes form or Describe the lifecycle of a star"
+        label="Enter your astrology query",
+        placeholder="e.g., I'm a Scorpio, born November 5, 1995. What's my future in career and love?"
     )
-    output_display = gr.Textbox(label="Astronomy Explanation", interactive=False, lines=10)
-    submit_button = gr.Button("Generate")
+    output_display = gr.Textbox(label="Your Fortune", interactive=False, lines=10)
+    submit_button = gr.Button("Predict")
     
     submit_button.click(
-        fn=live_astronomy_generator,
+        fn=live_astrology_generator,
         inputs=prompt_input,
         outputs=output_display
     )