ABEJA Qwen2.5 7b Japanese V0.1

🚀 ABEJA-Qwen2.5-7b-Japanese-v0.1

ABEJA-Qwen2.5-7b-Japanese-v0.1 is a model trained on Japanese based on Qwen/Qwen2.5-7B-Instruct. Instead of regular continued pre - training, it is a model trained through distillation learning based on abeja/ABEJA-Qwen2.5-32b-Japanese-v0.1. Post - Training is not carried out, and the instruction - following performance is improved by ChatVector (the difference vector between Qwen/Qwen2.5-7B-Instruct and Qwen/Qwen2.5-7B). For more details, please refer to ABEJA's tech blog.

🚀 Quick Start

ABEJA-Qwen2.5-7b-Japanese-v0.1 is a model that has undergone Japanese language training based on Qwen/Qwen2.5-7B-Instruct. Instead of regular continued pre - training, it uses distillation learning based on abeja/ABEJA-Qwen2.5-32b-Japanese-v0.1. Post - Training is not performed, and the instruction - following performance is enhanced by ChatVector (the difference vector between Qwen/Qwen2.5-7B-Instruct and Qwen/Qwen2.5-7B).

For more details, please refer to ABEJA's tech blog.

✨ Features

• Based on Qwen2.5: Built upon the Qwen/Qwen2.5-7B-Instruct base model.
• Distillation Learning: Utilizes distillation learning based on abeja/ABEJA-Qwen2.5-32b-Japanese-v0.1.
• Enhanced by ChatVector: Improves instruction - following performance using ChatVector.

💻 Usage Examples

Basic Usage

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "abeja/ABEJA-Qwen2.5-7b-Japanese-v0.1"

model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype="auto",
    device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained(model_name)

prompt = "人とAIが協調するためには？"
messages = [
    {"role": "system", "content": "You are Qwen, created by Alibaba Cloud. You are a helpful assistant."},
    {"role": "user", "content": prompt}
]
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True
)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)

generated_ids = model.generate(
    **model_inputs,
    max_new_tokens=512
)
generated_ids = [
    output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
]

response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]

print(response)

📄 License

This project is licensed under the Apache - 2.0 license.

👨‍💻 Developers

• Hiroshi Kiyota
• Keisuke Fujimoto
• Kentaro Nakanishi
• Kyo Hattori
• Shinya Otani
• Shogo Muranushi
• Takuma Kume
• Tomoki Manabe

(*) In alphabetical order