Awesome
bertology_sklearn: a python toolkit for natural language understanding with BERT related models
Key Features
- Easy Using
- Multiple NLP tasks
- BERTology models
- State of the art
Easy Using
-
Input Data
X
- shape:
[n_samples, max_seq_len]
- type:
list
|ndarray
|pd.Dataframe
- shape:
y
- shape:
[n_samples]
|[n_sampels, max_seq_len]
for NER |[n_sampels, n_classes]
for multi-label classification - type:
list
|ndarray
(numpy) |Dataframe
(pandas)
- shape:
-
Classes
BertologyClassifier
, used for customizing models for text classificationBertologyTokenClassifier
, used for customizing models for natural language recognition(NER)
-
Class methods
fit(X,y)
, used for fine-tuningpredict(X)
, used for predictingscore(X,y)
, used for scoring
-
Easy training
- early stopping
- k-fold cross validation
NLP tasks
- Name Entity Recognition
- Text Classification
- Binary Classes
- Multiple Classes
- Multi-Label Classes