Returns the most possible answers on a given question for QA models (BERT, Albert,
etc.).
The API expects a Bert based TFLite model with metadata containing the following
information:
- input_process_units for Wordpiece/Sentencepiece Tokenizer - Wordpiece Tokenizer can be
used for a MobileBert model,
Sentencepiece Tokenizer Tokenizer can be used for an Albert model.
- 3 input tensors with names "ids", "mask" and "segment_ids".
- 2 output tensors with names "end_logits" and "start_logits".
Inherited Method Summary
From class org.tensorflow.lite.task.core.BaseTaskApi
void |
checkNotClosed()
|
synchronized void |
close()
|
abstract void |
deinit(long arg0)
|
void |
finalize()
|
long |
getNativeHandle()
|
boolean |
isClosed()
|
From class java.lang.Object
Object
|
clone()
|
boolean |
|
void |
finalize()
|
final Class<?>
|
getClass()
|
int |
hashCode()
|
final void |
notify()
|
final void |
notifyAll()
|
String
|
toString()
|
final void |
wait(long arg0, int arg1)
|
final void |
wait(long arg0)
|
final void |
wait()
|
From interface java.io.Closeable
From interface java.lang.AutoCloseable
Public Methods
Answers question based on context, and returns a list of possible QaAnswer
s.
Could be empty if no answer was found from the given context.
Parameters
context |
context the question bases on |
question |
question to ask |
public static BertQuestionAnswerer
createAlbertQuestionAnswererFromFile
(Context context,
String modelPath,
String
sentencePieceModelPath)
Creates a BertQuestionAnswerer
instance with an Albert model and a sentence piece model file.
One suitable model is:
//tfhub.dev/tensorflow/lite-model/albert_lite_base/squadv1/1
Parameters
context |
android context |
modelPath |
file path to the Albert model. Note: The model should not be compressed |
sentencePieceModelPath |
file path to the sentence piece model file. Note: The model should not be
compressed |
public static BertQuestionAnswerer
createBertQuestionAnswererFromFile
(Context context,
String modelPath,
String
vocabPath)
Creates a BertQuestionAnswerer
instance with a Bert model and a vocabulary file.
One suitable model is: //tfhub.dev/tensorflow/lite-model/mobilebert/1/default/1
Parameters
context |
android context |
modelPath |
file path to the Bert model. Note: The model should not be compressed |
vocabPath |
file path to the vocabulary file. Note: The file should not be compressed |
Parameters
modelFile |
a File object of the
model |
Parameters
context |
android context |
modelPath |
file path to the model with metadata. Note: The model should not be
compressed |
Parameters
context |
android context |
modelPath |
file path to the model with metadata. Note: The model should not be
compressed |
options |
|
Parameters
modelFile |
a File object of the
model |
options |
|
Protected Methods
protected void deinit (long
nativeHandle)