-
Notifications
You must be signed in to change notification settings - Fork 4.3k
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Yuanchen Xu
committed
Jul 28, 2023
1 parent
ef4b99e
commit 649aca1
Showing
5 changed files
with
300 additions
and
15 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,87 @@ | ||
# Copyright 2023 lm-sys@FastChat | ||
# | ||
# Licensed under the Apache License, Version 2.0 (the "License"); | ||
# you may not use this file except in compliance with the License. | ||
# You may obtain a copy of the License at | ||
# | ||
# http://www.apache.org/licenses/LICENSE-2.0 | ||
# | ||
# Unless required by applicable law or agreed to in writing, software | ||
# distributed under the License is distributed on an "AS IS" BASIS, | ||
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | ||
# See the License for the specific language governing permissions and | ||
# limitations under the License. | ||
|
||
import dataclasses | ||
from enum import Enum, auto | ||
from typing import List | ||
|
||
|
||
class SeparatorStyle(Enum): | ||
ADD_EOS_TOKEN = auto() | ||
|
||
|
||
@dataclasses.dataclass | ||
class Conversation: | ||
system: str | ||
roles: List[str] | ||
messages: List[List[str]] | ||
offset: int | ||
sep_style: SeparatorStyle = SeparatorStyle.ADD_EOS_TOKEN | ||
sep: str = "</s>" | ||
|
||
skip_next: bool = False | ||
|
||
def get_prompt(self): | ||
if self.sep_style == SeparatorStyle.ADD_EOS_TOKEN: | ||
ret = self.system | ||
for role, message in self.messages: | ||
if message: | ||
ret += role + ": " + message + self.sep | ||
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) | ||
|
||
def dict(self): | ||
return { | ||
"system": self.system, | ||
"roles": self.roles, | ||
"messages": self.messages, | ||
"offset": self.offset, | ||
"sep": self.sep | ||
} | ||
|
||
|
||
conv = Conversation( | ||
system="A chat between a curious human and an artificial intelligence assistant. " | ||
"The assistant gives helpful, detailed, and polite answers to the human's questions.\n\n", | ||
roles=("Human", "Assistant"), | ||
messages=(), | ||
offset=0, | ||
sep_style=SeparatorStyle.ADD_EOS_TOKEN, | ||
sep="</s>", | ||
) | ||
|
||
default_conversation = conv |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
79 changes: 79 additions & 0 deletions
79
applications/Chat/examples/generate_conversation_dataset.py
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,79 @@ | ||
import argparse | ||
import json | ||
|
||
from datasets import load_dataset | ||
|
||
|
||
def generate_alpaca(): | ||
# We can convert dataset with the same format("instruction", "input", "output") as Alpaca into a one-round conversation. | ||
conversation_dataset = [] | ||
dataset = load_dataset("tatsu-lab/alpaca", split="train") | ||
|
||
instructions = dataset["instruction"] | ||
inputs = dataset["input"] | ||
outputs = dataset["output"] | ||
|
||
assert len(instructions) == len(inputs) == len(outputs) | ||
|
||
for idx in range(len(instructions)): | ||
human_utterance = instructions[idx] + "\n\n" + inputs[idx] if inputs[idx] else instructions[idx] | ||
human = {"from": "human", "value": human_utterance} | ||
|
||
gpt_utterance = outputs[idx] | ||
gpt = {"from": "gpt", "value": gpt_utterance} | ||
|
||
conversation = dict(type="instruction", language="English", dataset="Alpaca", conversations=[human, gpt]) | ||
conversation_dataset.append(conversation) | ||
|
||
return conversation_dataset | ||
|
||
|
||
def generate_sharegpt(): | ||
# ShareGPT data requires less processing. | ||
conversation_dataset = [] | ||
dataset = load_dataset("anon8231489123/ShareGPT_Vicuna_unfiltered", | ||
data_files="ShareGPT_V3_unfiltered_cleaned_split_no_imsorry.json", | ||
split="train") | ||
|
||
conversations = dataset["conversations"] | ||
|
||
for idx in range(len(conversations)): | ||
for conv in conversations[idx]: | ||
# We don't need markdown and text value. | ||
del conv["markdown"] | ||
del conv["text"] | ||
|
||
conversation = dict(type="conversation", | ||
language="Multilingual", | ||
dataset="ShareGPT", | ||
conversations=conversations[idx]) | ||
conversation_dataset.append(conversation) | ||
|
||
return conversation_dataset | ||
|
||
|
||
if __name__ == '__main__': | ||
parser = argparse.ArgumentParser() | ||
parser.add_argument('--dataset', | ||
type=str, | ||
default="All", | ||
choices=["Alpaca", "ShareGPT", "All"], | ||
help="which dataset to convert, All will combine Alpaca and ShareGPT") | ||
parser.add_argument('--save_path', type=str, default="dataset.json", help="path to save the converted dataset") | ||
args = parser.parse_args() | ||
|
||
conversation_dataset = [] | ||
|
||
if args.dataset == "Alpaca": | ||
conversation_dataset.extend(generate_alpaca()) | ||
elif args.dataset == "ShareGPT": | ||
conversation_dataset.extend(generate_sharegpt()) | ||
else: | ||
conversation_dataset.extend(generate_alpaca()) | ||
conversation_dataset.extend(generate_sharegpt()) | ||
|
||
for idx, sample in enumerate(conversation_dataset): | ||
sample["id"] = idx + 1 | ||
|
||
with open(args.save_path, mode='w') as f: | ||
json.dump(conversation_dataset, f, indent=4, default=str, ensure_ascii=False) |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters