Release v0.2.0

Updates:

• Remove transformer model
• Refactoring bert & albert modeling
• Refactoring loading pretrained bert & albert models

Release v0.1.4

Updates:

• Refactoring tokenizers using luozhouyang/naivenlp
• Refactoring load pretrained models, remove adapters module

Release v0.1.3

Updates:

• Add adapters to load pretrained models!

Release v0.1.2

Updates:

• Fixed mask dtype mismatch
• Add runners_test
• Fixed examples

Release v0.1.1

Updates:

• Add default value -1 to vocab_size, src_vocab_size, tgt_vocab_size
• Add assertion when vocab_size <= 0

Release v0.1.0

Updates:

• Refactoring Transformer model implementation
• Add BERT model implementation
• Add ALBERT model implementation
• Refactoring dataset builders
• Refactoring tokenizers
• Add runners to run models