Update app.py

app.py

@@ -1,23 +1,21 @@
-import gradio as gr
 import torch
-from transformers import AutoTokenizer, AutoModelForCausalLM
 from fastapi import FastAPI, Request
 from fastapi.responses import JSONResponse
-import urllib.parse
+from transformers import AutoTokenizer, AutoModelForCausalLM
 
 # Load model
 model_id = "microsoft/phi-2"
 tokenizer = AutoTokenizer.from_pretrained(model_id)
 model = AutoModelForCausalLM.from_pretrained(model_id)
 
-# Memory for users
+# Memory dict
 chat_history = {}
 
-# Format history
+# History formatter
 def format_context(history):
     return "".join([f"You: {u}\n𝕴 𝖆𝖒 𝖍𝖎𝖒: {b}\n" for u, b in history[-3:]])
 
-# FastAPI app
+# Create FastAPI app
 app = FastAPI()
 
 @app.get("/ai")
@@ -26,23 +24,17 @@ async def ai_chat(request: Request):
     user_input = query_params.get("query", "")
     user_id = query_params.get("user_id", "default")
 
-    # Get user history
+    # Pull history
     history = chat_history.get(user_id, [])
     prompt = format_context(history) + f"You: {user_input}\n𝕴 𝖆𝖒 𝖍𝖎𝖒:"
 
-    # Tokenize & run model
+    # Run model
     inputs = tokenizer(prompt, return_tensors="pt", return_attention_mask=True)
     outputs = model.generate(**inputs, max_new_tokens=100, pad_token_id=tokenizer.eos_token_id)
     reply = tokenizer.decode(outputs[0], skip_special_tokens=True).split("𝕴 𝖆𝖒 𝖍𝖎𝖒:")[-1].strip()
 
-    # Save memory
+    # Store memory
     history.append((user_input, reply))
     chat_history[user_id] = history[-10:]
 
-    return JSONResponse({"reply": reply})
-
-# Wrap with Gradio to serve
-app = gr.mount_gradio_app(app, gr.Interface(lambda x: x, "textbox", "textbox"))
-
-# Launch it
-gradio_app = gr.FastAPI(app)
+    return JSONResponse({"reply": reply})