Update main.py

main.py

@@ -2,20 +2,22 @@ from ctransformers import AutoModelForCausalLM
 from fastapi import FastAPI, Form
 from pydantic import BaseModel
 
-
+#Model loading
 llm = AutoModelForCausalLM.from_pretrained("zephyr-7b-beta.Q4_K_S.gguf",
 model_type='mistral',
 max_new_tokens = 1096,
-threads = 1,
+threads = 3,
 )
    
 
 #Pydantic object
 class validation(BaseModel):
     prompt: str
+    
 #Fast API
 app = FastAPI()
-#Function contain tranlater API, RAG API, OpenAI API
+
+#Zephyr completion
 @app.post("/llm_on_cpu")
 async def stream(item: validation):
     system_prompt = 'Below is an instruction that describes a task. Write a response that appropriately completes the request.'
@@ -23,11 +25,3 @@ async def stream(item: validation):
     user, assistant = "<|user|>", "<|assistant|>"
     prompt = f"{system_prompt}{E_INST}\n{user}\n{item.prompt.strip()}{E_INST}\n{assistant}\n"
     return llm(prompt)
-
-#def stream(user_prompt)
-# system_prompt = 'Below is an instruction that describes a task. Write a response that appropriately completes the request.'
-# E_INST = "</s>"
-# user, assistant = "<|user|>", "<|assistant|>"
-# prompt = f"{system_prompt}{E_INST}\n{user}\n{user_prompt.strip()}{E_INST}\n{assistant}\n"
-# for text in llm(prompt, stream=True):
-#     print(text, end="", flush=True)