Update app.py

app.py

@@ -3,13 +3,18 @@ from llama_cpp import Llama
 from typing import Generator
 import os
 
-# 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
+# Initialize model
+model_path = "AstroSage-8B-BF16.gguf"  # Downloaded from AstroMLab/AstroSage-8B-GGUF
+llm = Llama.from_pretrained(
+    repo_id="AstroMLab/AstroSage-8B-GGUF",
+    filename=model_path,
+    n_ctx=2048,  # Context length for prompts
+    n_threads=2  # Use 2 CPU cores
+)
 
 def generate_astrology_prediction(prompt: str) -> Generator[str, None, None]:
     """
-    Generates astrology-based fortune-telling predictions using AstroSage-8B-Q8_0-GGUF with streaming.
+    Generates astrology-based fortune-telling predictions using AstroSage-8B-BF16.gguf with streaming.
     """
     system_prompt = (
         "You are an expert astrologer, specializing in fortune-telling. Given a user prompt "
@@ -18,9 +23,9 @@ def generate_astrology_prediction(prompt: str) -> Generator[str, None, None]:
         "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."
     )
-    full_prompt = f"<|SYSTEM|> {system_prompt} <|USER|> {prompt} <|ASSISTANT|>"
+    full_prompt = f"<|SYSTEM|> {system_prompt}\n<|USER|> {prompt}\n<|ASSISTANT|>"
 
-    # Stream output from llama.cpp
+    # Stream output
     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: