What are the methods used for achieving compressed model?
by SMEBOOK (admin) · February 28, 2021
The following are some popular, heavily researched methods for achieving compressed models:
- Low-rank approximation and sparsity
- Knowledge distillation
- Neural Architecture Search (NAS)
- In practice, a complete model compression pipeline might integrate several of these approaches, as each comes with its own benefits and drawbacks.
Share it on social networks: Tweet Share
SMEBOOK is reinventing the management of tech companies’ assets by providing a matchmaking algorithm capable of recommending partnerships according to needs and interests.