feat: provide access method with api for 7B:int8

[streetycat]

you can access it with follow:

1. ui: input http://$host/, you can get a webpage.
2. api: POST http://$host/generate, with `body`:
     class RepetitionPenalty(BaseModel):
         range: Optional[int] = 1024
         slope: Optional[float] = 0
         value: Optional[float] = 1.15

     class GenerateParam(BaseModel):
         prompts: Union[List[str], str]
         max_gen_len: Optional[int] = 1024
         temperature: Optional[float] = 0.8
         top_p: Optional[float] = 0.95
         repetition_penalty: Optional[RepetitionPenalty] = RepetitionPenalty()

    you can get a json:
        {
              error: int,
              msg: "error messages",
              results: {results from llama}
        }