revert + enable api

app.py

@@ -1,7 +1,4 @@
 from huggingface_hub import InferenceClient
-from fastapi import FastAPI
-from pydantic import BaseModel
-
 
 import gradio as gr
 
@@ -91,20 +88,6 @@ additional_inputs=[
     )
 ]
 
-class Item(BaseModel):
-    prompt: str
-    history: list
-    system_prompt: str
-    temperature: float = 0.0
-    max_new_tokens: int = 1048
-    top_p: float = 0.15
-    repetition_penalty: float = 1.0
-app = FastAPI()
-
-@app.post("/generate/")
-async def generate_text(item: Item):
-    return {"response": generate(item.prompt, item.history, item.system_message, item.temperature, item.max_new_tokens, item.top_p, item.repetition_penalty)}
-
 
 
 gr.ChatInterface(
@@ -112,4 +95,4 @@ gr.ChatInterface(
     chatbot=gr.Chatbot(show_label=False, show_share_button=False, show_copy_button=True, likeable=True, layout="panel"),
     additional_inputs=additional_inputs,
     title="""mistralai/Mistral-7B-Instruct-v0.3"""
-).launch(show_api=False)
+).launch(show_api=True)