Create app.py

app.py

@@ -0,0 +1,58 @@
+import subprocess
+import time
+import requests
+import gradio as gr
+from huggingface_hub import hf_hub_download
+
+# =========================
+# 1. DOWNLOAD MODEL
+# =========================
+MODEL_PATH = hf_hub_download(
+    repo_id="unsloth/gemma-4-E4B-it-GGUF",
+    filename="gemma-4-E4B-it-IQ4_XS.gguf"
+)
+
+print("Model downloaded:", MODEL_PATH)
+
+# =========================
+# 2. START llama.cpp SERVER
+# =========================
+server = subprocess.Popen([
+    "/llama.cpp/build/bin/llama-server",
+    "-m", MODEL_PATH,
+    "-c", "2048",
+    "-t", "2",
+    "-ngl", "0",
+    "--host", "127.0.0.1",
+    "--port", "8080"
+])
+
+time.sleep(6)
+
+# =========================
+# 3. GRADIO CHAT FUNCTION
+# =========================
+def chat(message, history):
+    payload = {
+        "messages": [
+            {"role": "user", "content": message}
+        ]
+    }
+
+    r = requests.post(
+        "http://127.0.0.1:8080/v1/chat/completions",
+        json=payload,
+        timeout=300
+    )
+
+    return r.json()["choices"][0]["message"]["content"]
+
+# =========================
+# 4. GRADIO UI
+# =========================
+demo = gr.ChatInterface(chat)
+
+demo.launch(
+    server_name="0.0.0.0",
+    server_port=7860
+)