Update main.py

main.py

@@ -1,21 +1,21 @@
 from fastapi import FastAPI, Form
 from fastapi.middleware.cors import CORSMiddleware
 from pydantic import BaseModel
-from transformers import AutoModelForCausalLM
+from ctransformers import AutoModelForCausalLM
 
-# Model loading
-llm = AutoModelForCausalLM.from_pretrained(
-    "zephyr-7b-beta.Q4_K_S.gguf",
-    model_type='mistral',
-    max_new_tokens=1096,
-    threads=3,
+#Model loading
+llm = AutoModelForCausalLM.from_pretrained("zephyr-7b-beta.Q4_K_S.gguf",
+model_type='mistral',
+max_new_tokens = 1096,
+threads = 3,
 )
+   
 
-# Pydantic object
-class Validation(BaseModel):
+#Pydantic object
+class validation(BaseModel):
     prompt: str
-
-# Fast API
+    
+#Fast API
 app = FastAPI()
 
 # Set up CORS
@@ -32,11 +32,11 @@ app.add_middleware(
     allow_headers=["*"],
 )
 
-# Zephyr completion
+#Zephyr completion
 @app.post("/llm_on_cpu")
-async def stream(item: Validation):
+async def stream(item: validation):
     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 = "<|user|>", "<|assistant|>"
     prompt = f"{system_prompt}{E_INST}\n{user}\n{item.prompt.strip()}{E_INST}\n{assistant}\n"
     return llm(prompt)