Update app.py

app.py

@@ -1,15 +1,17 @@
 import gradio as gr
 from llama_cpp import Llama
 
-print("Downloading model")
+
 llm = Llama.from_pretrained(
-    repo_id="bartowski/gemma-2-2b-it-abliterated-GGUF",
-    filename="gemma-2-2b-it-abliterated-IQ4_XS.gguf",
+    repo_id="bartowski/Phi-3.5-mini-instruct-GGUF",
+    filename="Phi-3.5-mini-instruct-Q4_K_M.gguf",
     numa=True,
-    n_ctx=4096,
+    # flash_attn=True,
+    # n_gpu_layers=-1,
+    n_batch=1024,
+    n_ctx=4095,
 )
 
-
 def respond(prompt: str):
     stream = llm.create_chat_completion(stream=True, messages=[{"role": "user", "content": prompt}])