Unleashing the Potential of PIM: Accelerating Large Batched Inference of Transformer-Based Generative Models.
Jaewan ChoiJaehyun ParkKwanhee KyungNam Sung KimJung Ho AhnPublished in: IEEE Comput. Archit. Lett. (2023)
Keyphrases
- generative model
- probabilistic model
- parameter learning
- boltzmann machine
- representational power
- discriminative learning
- mixture model
- prior knowledge
- em algorithm
- discriminative models
- bayesian framework
- posterior probability
- conditional random fields
- generative and discriminative models
- bayesian inference
- pitman yor process
- bayesian networks
- semi supervised
- probabilistic inference
- generative process
- hierarchical hidden markov models
- latent dirichlet allocation
- object categories
- object recognition
- topic models
- image processing
- information retrieval