Prismer: A Vision-Language Model with An Ensemble of Experts
Abstract
Recent vision-language models have shown impressive multi-modal generation capabilities. However, typically they require training huge models on massive datasets. As a more scalable alternative, we introduce Prismer, a data- and parameter-efficient vision-language model that leverages an ensemble of domain experts. Prismer only requires training of a small number of components, with the majority of network weights inherited from readily-available, pre-trained domain experts, and kept frozen during training. By leveraging experts from a wide range of domains, we show that Prismer can efficiently pool this expert knowledge and adapt it to various vision-language reasoning tasks. In our experiments, we show that Prismer achieves fine-tuned and few-shot learning performance which is competitive with current state-of-the-art models, whilst requiring up to two orders of magnitude less training data. Code is available at https://github.com/NVlabs/prismer.
Attached Files
Submitted - 2303.02506.pdf
Files
Name | Size | Download all |
---|---|---|
md5:b91d6904e7ff02751e2fd6f3c98e1203
|
5.5 MB | Preview Download |
Additional details
- Eprint ID
- 120079
- Resolver ID
- CaltechAUTHORS:20230316-153658096
- Created
-
2023-03-16Created from EPrint's datestamp field
- Updated
-
2023-03-16Created from EPrint's last_modified field