Mihir Kale
Researcher / author (OpenAlex)
Facts
1 property
🔗 Connections
Author of 6
ByT5: Towards a Token-Free Future with Pre-trained Byte-to-Byte Models,
Improving Compositional Generalization with Self-Training for Data-to-Text Generation,
mT5: A Massively Multilingual Pre-trained Text-to-Text Transformer,
nmT5 - Is parallel data still relevant for pre-training massively multilingual language models?,
Template Guided Text Generation for Task-Oriented Dialogue,
Text-to-Text Pre-Training for Data-to-Text Tasks
← Authored by 6
ByT5: Towards a Token-Free Future with Pre-trained Byte-to-Byte Models,
Improving Compositional Generalization with Self-Training for Data-to-Text Generation,
mT5: A Massively Multilingual Pre-trained Text-to-Text Transformer,
nmT5 - Is parallel data still relevant for pre-training massively multilingual language models?,
Template Guided Text Generation for Task-Oriented Dialogue,
Text-to-Text Pre-Training for Data-to-Text Tasks
Affiliated with 1
Press Enter · cited answer in seconds
0 sources
Summary
Mihir Kale is a human[1].