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。

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

* rename preprocess.py into utils.py, because nothing about preprocess in it
* anything in loader/ and saver/ is moved directly into io/
* corresponding unit tests are moved to /test/io
* delete fastnlp.py, because we have new and better APIs
* rename Biaffine_parser/run_test.py to Biaffine_parser/main.py; Otherwise, test will fail.
* A looooooooooot of ancient codes to be refined...........

- 修复processor的bug
- 更新core/的若干组件, 去除batch的冗余参数
- CRF有个打字错误？已修复
- 更新pos tag 训练脚本

# Conflicts:
#	fastNLP/api/pipeline.py
#	fastNLP/api/pos_tagger.py
#	fastNLP/api/processor.py
#	fastNLP/modules/decoder/CRF.py

- 在api中创建converter.py
- Pipeline添加初始化方法，方便一次性添加processors
- 删除pos_tagger.py
- 优化整体code style

- update predictor.py to remove unused methods
- update model_loader.py & model_saver.py to support entire model saving & loading
- update pos tagging training script