* Add embedding seq-cls finetune demo and update api

* Update docs of pad_sequence and trunc_sequence

* Reset metric value at training and validation step

* Add text processing of token-cls task in predict method

* Add token-cls task for transformer modules

* Fix numpy dtype mismatch in windows

* Update README.md

* Update token-cls task for ernie_tiny

* Update token-cls task for ernie_tiny

* Update token-cls task for other Transformer modules

* Update README.md of modules and token-cls demo

* Add chunk_scheme arg check in SeqLabelingDataset

* Update ChunkEvaluator and paddlenlp requirement

* Update README.md

* Update token-cls demo

* Add MSRA dataset

* Add ernie_tiny seq-cls task case

* Update README.md