57 lines
1.9 KiB
Markdown
57 lines
1.9 KiB
Markdown
|
---
|
||
|
|
language:
|
||
|
|
- ja
|
||
|
|
- en
|
||
|
|
license: other
|
||
|
|
library_name: transformers
|
||
|
|
license_name: gemma-terms-of-use
|
||
|
|
license_link: https://www.kaggle.com/models/google/gemma/license/consent
|
||
|
|
inference: false
|
||
|
|
---
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
# はじめに
|
||
|
|
GoogleのGemma-2Bを日本語で使えるように継続事前学習を施した、商用利用可能なベースモデルです。
|
||
|
|
小型なのでスマホや家電などに向いています。**ただし、Instruction tuningが困難な可能性があります。**
|
||
|
|
|
||
|
|
[Colabで試す](https://colab.research.google.com/drive/1g4739fY-n4RiS9-BwwiPJI92sRahxGuO?usp=sharing)
|
||
|
|
|
||
|
|
mmngaさんが作った[軽量版](https://huggingface.co/mmnga/alfredplpl-suzume-poc-gguf)を[Colabで試す](https://colab.research.google.com/drive/18vsNSpWx0Ydp3KYTnUqlChTct_OvuToO?usp=sharing)
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
# Usage
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
```python
|
||
|
|
from transformers import AutoTokenizer, AutoModelForCausalLM
|
||
|
|
|
||
|
|
tokenizer = AutoTokenizer.from_pretrained("alfredplpl/suzume-poc")
|
||
|
|
model = AutoModelForCausalLM.from_pretrained("alfredplpl/suzume-poc")
|
||
|
|
|
||
|
|
input_text = """人工知能とは"""
|
||
|
|
input_ids = tokenizer(input_text, return_tensors="pt")
|
||
|
|
|
||
|
|
outputs = model.generate(**input_ids,max_new_tokens=64)
|
||
|
|
print(tokenizer.decode(outputs[0]))
|
||
|
|
|
||
|
|
```
|
||
|
|
|
||
|
|
## Result
|
||
|
|
|
||
|
|
```bash
|
||
|
|
<bos>人工知能とは、「人工生物」に基づいた知能、意図、計画、感情などの概念を構築し、人間のような思考・意思・判断・判断などを備えた機体とする試み。
|
||
|
|
本稿では、人工知能やAIに関する動向をまとめる。なお、「人工知能」と
|
||
|
|
```
|
||
|
|
|
||
|
|
|
||
|
|
# Base model
|
||
|
|
- google/gemma-2b
|
||
|
|
|
||
|
|
# Dataset for continual pretraining
|
||
|
|
- izumi-lab/wikipedia-ja-20230720
|
||
|
|
|
||
|
|
# How to make this model
|
||
|
|
[<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl)
|