TextClassifier

Performs classification on text.

This API expects a TFLite model with (optional) TFLite Model Metadata that contains the mandatory (described below) input tensors, output tensor, and the optional (but recommended) label items as AssociatedFiles with type TENSOR_AXIS_LABELS per output classification tensor.

Metadata is required for models with int32 input tensors because it contains the input process unit for the model's Tokenizer. No metadata is required for models with string input tensors.

• Input tensors
  • Three input tensors (kTfLiteInt32) of shape [batch_size x bert_max_seq_len] representing the input ids, mask ids, and segment ids. This input signature requires a Bert Tokenizer process unit in the model metadata.
  • Or one input tensor (kTfLiteInt32) of shape [batch_size x max_seq_len] representing the input ids. This input signature requires a Regex Tokenizer process unit in the model metadata.
  • Or one input tensor (kTfLiteString) that is shapeless or has shape [1] containing the input string.
• At least one output tensor (kTfLiteFloat32/kBool) with:
  • N classes and shape [1 x N]
  • optional (but recommended) label map(s) as AssociatedFile-s with type TENSOR_AXIS_LABELS, containing one label per line. The first such AssociatedFile (if any) is used to fill the class_name field of the results. The display_name field is filled from the AssociatedFile (if any) whose locale matches the display_names_locale field of the TextClassifierOptions used at creation time ("en" by default, i.e. English). If none of these are available, only the index field of the results will be filled.