ELYZA-japanese-CodeLlama-7b-instruct-GPTQ-calib-ja-1k

elyzaさんが公開しているELYZA-japanese-CodeLlama-7b-instructを 日本語のキャリブレーションセットで生成したGPTQモデルになります。

キャリブレーションセットはizumi-lab/wikipedia-ja-20230720から、 1kほどランダムサンプリングしたものと、
ELYZA-tasks-100のinput/outputを計200ほど追加しています。
mmnga/wikipedia-ja-20230720-1k

他のモデルはこちら

通常版: llama2に日本語のデータセットで学習したモデル
mmnga/ELYZA-japanese-Llama-2-7b-gguf
mmnga/ELYZA-japanese-Llama-2-7b-instruct-gguf

Fast版 日本語の語彙を追加してトークンコストを減らし、1.8倍高速化したモデル
mmnga/ELYZA-japanese-Llama-2-7b-fast-gguf
mmnga/ELYZA-japanese-Llama-2-7b-fast-instruct-gguf

Codellama版 GGUF
mmnga/ELYZA-japanese-CodeLlama-7b-gguf
mmnga/ELYZA-japanese-CodeLlama-7b-instruct-gguf

Codellama版 GPTQ
mmnga/ELYZA-japanese-CodeLlama-7b-instruct-GPTQ-calib-ja-1k

Usage

pip install auto-gptq[triton]==0.4.2 transformers==4.34.1

import torch
from auto_gptq import AutoGPTQForCausalLM, BaseQuantizeConfig
from transformers import AutoTokenizer , AutoModelForCausalLM

if torch.cuda.is_available():
    device_name = torch.cuda.get_device_name(0)

model_name_or_path = "mmnga/ELYZA-japanese-CodeLlama-7b-instruct-GPTQ-calib-ja-1k"

# Tokenizer
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, trust_remote_code=True)

# Model
model = AutoGPTQForCausalLM.from_quantized(model_name_or_path, use_safetensors=True, device="cuda:0", use_triton=("A100" in device_name))

# Your test prompt
B_INST, E_INST = "[INST]", "[/INST]"
B_SYS, E_SYS = "<<SYS>>\n", "\n<</SYS>>\n\n"
DEFAULT_SYSTEM_PROMPT = "あなたは誠実で優秀な日本人のアシスタントです。"
text = "エラトステネスの篩についてサンプルコードを示し、解説してください。"

prompt = "{b_inst} {system}{prompt} {e_inst} ".format(
    b_inst=B_INST,
    system=f"{B_SYS}{DEFAULT_SYSTEM_PROMPT}{E_SYS}",
    prompt=text,
    e_inst=E_INST,
)

print(tokenizer.decode(model.generate(**tokenizer(prompt, return_tensors="pt").to(model.device), max_length=1024)[0]))

Licence

Llama 2 is licensed under the LLAMA 2 Community License, Copyright (c) Meta Platforms, Inc. All Rights Reserved.

引用 Citations

@misc{elyzacodellama2023, 
      title={ELYZA-japanese-CodeLlama-7b}, 
      url={https://huggingface.co/elyza/ELYZA-japanese-CodeLlama-7b}, 
      author={Akira Sasaki and Masato Hirakawa and Shintaro Horie and Tomoaki Nakamura},
      year={2023},
}

@misc{rozière2023code,
      title={Code Llama: Open Foundation Models for Code}, 
      author={Baptiste Rozière and Jonas Gehring and Fabian Gloeckle and Sten Sootla and Itai Gat and Xiaoqing Ellen Tan and Yossi Adi and Jingyu Liu and Tal Remez and Jérémy Rapin and Artyom Kozhevnikov and Ivan Evtimov and Joanna Bitton and Manish Bhatt and Cristian Canton Ferrer and Aaron Grattafiori and Wenhan Xiong and Alexandre Défossez and Jade Copet and Faisal Azhar and Hugo Touvron and Louis Martin and Nicolas Usunier and Thomas Scialom and Gabriel Synnaeve},
      year={2023},
      eprint={2308.12950},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}

@misc{touvron2023llama,
      title={Llama 2: Open Foundation and Fine-Tuned Chat Models}, 
      author={Hugo Touvron and Louis Martin and Kevin Stone and Peter Albert and Amjad Almahairi and Yasmine Babaei and Nikolay Bashlykov and Soumya Batra and Prajjwal Bhargava and Shruti Bhosale and Dan Bikel and Lukas Blecher and Cristian Canton Ferrer and Moya Chen and Guillem Cucurull and David Esiobu and Jude Fernandes and Jeremy Fu and Wenyin Fu and Brian Fuller and Cynthia Gao and Vedanuj Goswami and Naman Goyal and Anthony Hartshorn and Saghar Hosseini and Rui Hou and Hakan Inan and Marcin Kardas and Viktor Kerkez and Madian Khabsa and Isabel Kloumann and Artem Korenev and Punit Singh Koura and Marie-Anne Lachaux and Thibaut Lavril and Jenya Lee and Diana Liskovich and Yinghai Lu and Yuning Mao and Xavier Martinet and Todor Mihaylov and Pushkar Mishra and Igor Molybog and Yixin Nie and Andrew Poulton and Jeremy Reizenstein and Rashi Rungta and Kalyan Saladi and Alan Schelten and Ruan Silva and Eric Michael Smith and Ranjan Subramanian and Xiaoqing Ellen Tan and Binh Tang and Ross Taylor and Adina Williams and Jian Xiang Kuan and Puxin Xu and Zheng Yan and Iliyan Zarov and Yuchen Zhang and Angela Fan and Melanie Kambadur and Sharan Narang and Aurelien Rodriguez and Robert Stojnic and Sergey Edunov and Thomas Scialom},
      year={2023},
      eprint={2307.09288},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}