Create app.py

app.py

@@ -0,0 +1,20 @@
+from fastapi import FastAPI
+from transformers import AutoModelForCausalLM, AutoTokenizer
+import torch
+
+app = FastAPI()
+
+MODEL_NAME = "deepseek-ai/deepseek-llm-7b"
+tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
+model = AutoModelForCausalLM.from_pretrained(MODEL_NAME, torch_dtype=torch.float16, device_map="auto")
+
+@app.get("/")
+def root():
+    return {"message": "DeepSeek API is running!"}
+
+@app.post("/summarize/")
+async def summarize(input_text: str):
+    inputs = tokenizer(input_text, return_tensors="pt").to("cuda")
+    output = model.generate(**inputs, max_new_tokens=100)
+    summary = tokenizer.decode(output[0], skip_special_tokens=True)
+    return {"summary": summary}