Django-based text labeling platform

Abstract

The invention relates to a Django-based text labeling platform, which comprises an algorithm module, a project module, a user module and a label module, the user module is used for storing user information and performing login verification; the label module adds, deletes and modifies labels of projects, different labels are set for different projects, and the same project label cannot be repeatedand is used for labeling according to label types; and the algorithm module selects an algorithm in a project through the project module according to a packaging algorithm program, pre-annotation model training is carried out, the project module transmits annotated data of the project to the algorithm module, after a pre-annotation model is trained, pre-annotation is carried out through the project module, and unannotated data in the project is transmitted to the pre-annotation model. The algorithm module stores the labeled data into a database, the project module pre-labels all files which donot reach the standard in a project, and the pre-labeled data is checked on a system interface after pre-labeling is completed.

Description

technical field [0001] The invention relates to artificial intelligence natural language technology, in particular to a Django-based text labeling platform. Background technique [0002] In recent years, with the rapid development of the field of artificial intelligence, natural language processing has received widespread attention as an important research direction of artificial intelligence. Natural language processing technology is mainly used to solve problems such as sequence labeling and classification, most of which belong to supervised learning, and need to use labeled data to train corresponding models. The data labeling process requires a lot of manpower to classify data through manual judgment. Traditional manual labeling is done by distributing data to individual labelers, but there are the following defects: 1. The data format marked by different personnel may be biased, which is not convenient for unified processing; 2. The labeling speed is slow, and it is pr...

AI Technical Summary

Problems solved by technology

Traditional manual labeling is done by distributing data to individual labelers, but there are the following defects: 1. The data format marked by different personnel may be biased, which is not convenient for unified processing; 2. The labeling speed is slow and prone to mislabeling 3. Lack of collaboration between personnel and no cross-validation; 4. Low management efficiency, unable to adjust projects in real time; 5. No pre-labeling function, which consumes more manpower

Images