v

app.py

@@ -7,7 +7,7 @@ def start_server():
     os.system("uvicorn server:app --port 8080 --host 0.0.0.0 --workers 1")
     os.environ["SSTART"] = "1"
     
-if os.environ.get('SSTART') is None:
+if os.environ.get('SSTART') != "1":
     start_server()     
 
 st.set_page_config(

server.py

@@ -7,10 +7,11 @@ app = FastAPI()
 model_name = 'facebook/incoder-1B'
 tokenizer = AutoTokenizer.from_pretrained(model_name)
 model = AutoModelForCausalLM.from_pretrained(model_name, low_cpu_mem_usage=True)
+print('load ok')
 
 @app.get("/")
 def read_root(input_text, max_length, top_p, top_k, num_beams, temperature, repetition_penalty):
-    inpt = model_name.encode(input_text, return_tensors="pt")
+    inpt = tokenizer.encode(input_text, return_tensors="pt")
     out = model.generate(inpt, max_length=max_length, top_p=top_p, top_k=top_k, temperature=temperature, num_beams=num_beams, repetition_penalty=repetition_penalty)
     res = tokenizer.decode(out[0])
     return {res}