Fix bug with weights.

main.py

@@ -2,17 +2,18 @@ from ctransformers import AutoModelForCausalLM
 from fastapi import FastAPI
 from pydantic import BaseModel
 
-
-llm = AutoModelForCausalLM.from_pretrained("zephyr-7b-beta.Q4_K_S.gguf",
-model_type='mistral',
-max_new_tokens = 1096,
-threads = 3,
-)
+#
+# 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):
     prompt: str
 #Fast API
+
 app = FastAPI()
 
 @app.post("/llm_on_cpu")
@@ -21,4 +22,5 @@ async def stream(item: validation):
     E_INST = "</s>"
     user, assistant = "<|user|>", "<|assistant|>"
     prompt = f"{system_prompt}{E_INST}\n{user}\n{item.prompt}{E_INST}\n{assistant}\n"
-    return llm(prompt)
+    # return llm(prompt)
+    return prompt