quick_start.md

Quick Start

This tutorial introduces how to install AdaSeq and use it train a model.

• 1. Requirements and Installation
  • a. Installation from source
  • b. Installation via pip
• 2. Example Usage
  • a. Usage by code
  • b. Usage via command-line tool

1. Requirements and Installation

AdaSeq project is based on Python version >= 3.7 and PyTorch version >= 1.8.

1.a Installation from source

git clone https://github.com/modelscope/adaseq.git
cd adaseq
pip install -r requirements.txt -f https://modelscope.oss-cn-beijing.aliyuncs.com/releases/repo.html

1.b Installation via pip

pip install adaseq

2. Example Usage

Let's train a Bert-CRF model for NER on the resume dataset as an example. All you need to do is to write a configuration file and use it to run a command.

We've already prepared a configuration file resume.yaml for you. Try it!

2.a Usage by code

2.a.1 Train a model

python scripts/train.py -c examples/bert_crf/configs/resume.yaml

(b) Test a model

python scripts/test.py -w ${checkpoint_dir}

2.b Usage via command-line tool

2.b.1 Train a model

adaseq train -c examples/bert_crf/configs/resume.yaml

2.b.2 Test a model

adaseq test -w ${checkpoint_dir}