Not-WizardLM-2-7B

amazingvince/Not-WizardLM-2-7B

<a href="https://colab.research.google.com/gist/pszemraj/d3d74ceab942722b49188606785e2bfd/not-wizardlm-2-7b-inference.ipynb"> <img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/> </a>

Included is code ripped from fastchat with the expected chat templating.

Also wiz.pdf is a pdf of the github blog showing the apache 2 release. Link to wayback machine included: https://web.archive.org/web/20240415221214/https://wizardlm.github.io/WizardLM2/

example

import dataclasses
from enum import auto, Enum
from typing import List, Tuple, Any


class SeparatorStyle(Enum):
    """Different separator style."""
    SINGLE = auto()
    TWO = auto()


@dataclasses.dataclass
class Conversation:
    """A class that keeps all conversation history."""
    system: str
    roles: List[str]
    messages: List[List[str]]
    offset: int
    sep_style: SeparatorStyle = SeparatorStyle.SINGLE
    sep: str = "###"
    sep2: str = None

    # Used for gradio server
    skip_next: bool = False
    conv_id: Any = None

    def get_prompt(self):
        if self.sep_style == SeparatorStyle.SINGLE:
            ret = self.system
            for role, message in self.messages:
                if message:
                    ret += self.sep + " " + role + ": " + message
                else:
                    ret += self.sep + " " + role + ":"
            return ret
        elif self.sep_style == SeparatorStyle.TWO:
            seps = [self.sep, self.sep2]
            ret = self.system + seps[0]
            for i, (role, message) in enumerate(self.messages):
                if message:
                    ret += role + ": " + message + seps[i % 2]
                else:
                    ret += role + ":"
            return ret
        else:
            raise ValueError(f"Invalid style: {self.sep_style}")

    def append_message(self, role, message):
        self.messages.append([role, message])

    def to_gradio_chatbot(self):
        ret = []
        for i, (role, msg) in enumerate(self.messages[self.offset:]):
            if i % 2 == 0:
                ret.append([msg, None])
            else:
                ret[-1][-1] = msg
        return ret

    def copy(self):
        return Conversation(
            system=self.system,
            roles=self.roles,
            messages=[[x, y] for x, y in self.messages],
            offset=self.offset,
            sep_style=self.sep_style,
            sep=self.sep,
            sep2=self.sep2,
            conv_id=self.conv_id)

    def dict(self):
        return {
            "system": self.system,
            "roles": self.roles,
            "messages": self.messages,
            "offset": self.offset,
            "sep": self.sep,
            "sep2": self.sep2,
            "conv_id": self.conv_id,
        }



conv = Conversation(
    system="A chat between a curious user and an artificial intelligence assistant. "
           "The assistant gives helpful, detailed, and polite answers to the user's questions.",
    roles=("USER", "ASSISTANT"),
    messages=[],
    offset=0,
    sep_style=SeparatorStyle.TWO,
    sep=" ",
    sep2="</s>",
)

conv.append_message(conv.roles[0], "Why would Microsoft take this down?")
conv.append_message(conv.roles[1], None)
prompt = conv.get_prompt()

inputs = tokenizer(prompt, return_tensors="pt").to(model.device)

result = model.generate(**inputs, max_new_tokens=1000)
generated_ids = result[0]
generated_text = tokenizer.decode(generated_ids, skip_special_tokens=True)
print(generated_text)