2. metric增加SpanFMetric，可以用于计算sequence labelling的performance
3. 分词复现任务根据新版接口做了部分调整。

* refine code style

* update fastNLP.core.rst
* refine all docstrings in core/

[new] Add a loader for conll2003 dataset

[core] improve trainer, loss and metrics

[test]Updates

* fix failed tests

[Doc] update

2. 移动Tester中device获取位置
3. trainer中
    (1) train()返回一个dict，并重新加载最佳模型
    (2) check_code时返回batch_x中每个数据的形状等信息

update

update README.md

update requirements in README.md

move the first table to the right place

update requirements

FastNLP v0.2

* remove conflicts
* all tests passed

* add DataSet.get_field(), to fetch a FieldArray based on its name
* remove old tutorials & add new tutorials

* refine code style
* fix tests
* add a new tutorial

* add code comments
* merge *_saver.py & *_loader.py in io/
* (ancient codes) rename Loss into LossFromTorch

* clean up unused codes

* remove unused codes in metrics.py
* add tests for DataSet
* add tests for FieldArray
* add tests for metrics.py
* fix predictor, add tests for predictor
* fix bucket sampler, add tests for bucket sampler