ngraph.ctc_greedy_decoder_seq_len¶
-
ngraph.ctc_greedy_decoder_seq_len(data: Union[_pyngraph.Node, int, float, numpy.ndarray], sequence_length: Union[_pyngraph.Node, int, float, numpy.ndarray], blank_index: Optional[Union[_pyngraph.Node, int, float, numpy.ndarray]] = None, merge_repeated: bool = True, classes_index_type: str = 'i32', sequence_length_type: str = 'i32', name: Optional[str] = None) → _pyngraph.Node¶ Return a node which performs CTCGreedyDecoderSeqLen.
- Parameters
data – The input 3D tensor. Shape: [batch_size, seq_length, num_classes]
sequence_length – Input 1D tensor with sequence length. Shape: [batch_size]
blank_index – Scalar or 1D tensor with specifies the class index to use for the blank class. Optional parameter. Default value is num_classes-1.
- Return:
The new node which performs CTCGreedyDecoderSeqLen.