NGC | Catalog
CatalogCollectionsTLT/Jarvis - Intent Detection & Slot Tagging

TLT/Jarvis - Intent Detection & Slot Tagging

For contents of this collection and more information, please view on a desktop device.
Logo for TLT/Jarvis - Intent Detection & Slot Tagging

Description

This collection contains models and notebooks for Intent detection & Slot tagging training and deployment with TLT and Jarvis respectively

Curator

NVIDIA

Modified

July 28, 2021
Containers
Helm Charts
Models
Resources

Intent Detection and Slot Tagging Collection

Overview

This page contains the information about the Intent Detection and Slot tagging collection with TLT. The models in this collection can be used to detect Intents and to tag Slots (Entities) in user queries in chat bots or voice assistant systems.

These models are usually data specific and will recognize specific intents and slots that were presented in a training dataset.

Model Architecture

These are pretrained Bert-like models with 2 linear classifier heads on the top of it. One for classifying an intent of the query on the output of the first token [CLS] and another for classifying slots for every other token of the query.

These models were trained with the combined loss function on the Intent and Slot classification task for a training dataset.

Available Models

For instructions on how to use a model, please see its corresponding model card page:

  • Intent Slot Classification (weather chat bot)

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

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.