new code

Dockerfile

@@ -1,5 +1,5 @@
 # Use an official Python runtime as a parent image
-FROM python:3.11
+FROM python:3.13
 
 # Set the working directory in the container
 WORKDIR /app

README.md

@@ -3,8 +3,7 @@ title: LLM Model
 emoji: 🤖
 colorFrom: indigo
 colorTo: purple
-sdk: gradio
-sdk_version: "4.36.0"
+sdk: docker
 app_file: app.py
 pinned: false
 ---

app.py

@@ -4,28 +4,32 @@ from transformers import AutoTokenizer, AutoModelForCausalLM
 import torch
 
 
+
 app = FastAPI()
 
 class ChatRequest(BaseModel):
     message: str
 
+
+# Load DeepSeek model (small one for local use)
+# Try bigger models if you have a GPU with >12GB VRAM
+model_name = "deepseek-ai/deepseek-coder-1.3b-instruct"
+
+print("Loading model... this may take a minute ⏳")
+global tokenizer
+global model
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForCausalLM.from_pretrained(
+    model_name,
+    torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32,
+    device_map="auto"
+)
+print("Model loaded ✅")
+
 @app.get("/")
 def root():
     
-    # Load DeepSeek model (small one for local use)
-    # Try bigger models if you have a GPU with >12GB VRAM
-    model_name = "deepseek-ai/deepseek-coder-1.3b-instruct"
-
-    print("Loading model... this may take a minute ⏳")
-    global tokenizer
-    global model
-    tokenizer = AutoTokenizer.from_pretrained(model_name)
-    model = AutoModelForCausalLM.from_pretrained(
-        model_name,
-        torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32,
-        device_map="auto"
-    )
-    print("Model loaded ✅")
+   
 
     return {"status": "ok"}