BaseTextRecogPostprocessor¶
- class mmocr.models.textrecog.BaseTextRecogPostprocessor(dictionary, max_seq_len=40, ignore_chars=['padding'], **kwargs)[source]¶
Base text recognition postprocessor.
- Parameters
dictionary (dict or
Dictionary
) – The config for Dictionary or the instance of Dictionary.max_seq_len (int) – max_seq_len (int): Maximum sequence length. The sequence is usually generated from decoder. Defaults to 40.
ignore_chars (list[str]) – A list of characters to be ignored from the final results. Postprocessor will skip over these characters when converting raw indexes to characters. Apart from single characters, each item can be one of the following reversed keywords: ‘padding’, ‘end’ and ‘unknown’, which refer to their corresponding special tokens in the dictionary.
- Return type
- get_single_prediction(probs, data_sample=None)[source]¶
Convert the output probabilities of a single image to index and score.
- Parameters
probs (torch.Tensor) – Character probabilities with shape \((T, C)\).
data_sample (TextRecogDataSample) – Datasample of an image.
- Returns
Index and scores per-character.
- Return type