TLT/Riva - Text Classification
Text Classification
Overview
This page contains the information about the Text Classification collection with TLT. Text classification models can be used for text classification problems such as sentiment analysis or domain/intent detection for dialogue systems. A text sequence is given to such models as input and the models predict a label for it.
These models are usually data specific and will recognize specific text categories or query domains that were presented in a training dataset.
Model Architecture
All the models have a simple and very effective architecture based on BERT-like models. Text classification models consists of two main modules:
- A encoder module which is a pre-trained BERT-like models such as BERT, RoBEERTa or Megatron.
- A decoder module which is an MLP classifier on the output of the first token [CLS].
Available Models
For instructions on how to use a model, please see its corresponding model card page:
License
License to use these models is covered by the Model EULA. By downloading the model checkpoints, you accept the terms and conditions of these licenses.
Suggested Reading
- More information about the Transfer Learning Toolkit can be found at the NVIDIA Developer Zone: https://developer.nvidia.com/transfer-learning-toolkit
- Read the TLT getting Started guide and release notes.
- More information about the experiment spec files can be found in the TLT User Guide
Ethical AI
NVIDIA’s platforms and application frameworks enable developers to build a wide array of AI applications. Consider potential algorithmic bias when choosing or creating the models being deployed. Work with the model’s developer to ensure that it meets the requirements for the relevant industry and use case; that the necessary instruction and documentation are provided to understand error rates, confidence intervals, and results; and that the model is being used under the conditions and in the manner intended.