5 papers
Search Your Block Floating Point Scales!
Tanmaey Gupta, Hayden Prairie, Xiaoxia Wu +10
Quantization has emerged as a standard technique for accelerating inference for generative models by enabling faster low-precision computations and reduced memory transfers. Recent…
Chipmunk: Training-Free Acceleration of Diffusion Transformers with Dynamic Column-Sparse Deltas
Austin Silveria, Soham V. Govande, Daniel Y. Fu
Diffusion Transformers (DiTs) have achieved state-of-the-art performance in high-quality image and video generation but incur substantial compute cost at inference. A common observ…
Projection: A Mixed-Initiative Research Process
Austin Silveria
Communication of dense information between humans and machines is relatively low bandwidth. Many modern search and recommender systems operate as machine learning black boxes, givi…
GraphQL Live Querying with DynamoDB
Austin Silveria
We present a method of implementing GraphQL live queries at the database level. Our DynamoDB simulation in Go mimics a distributed key-value store and implements live queries to ex…
Exploratory Search with Sentence Embeddings
Austin Silveria
Exploratory search aims to guide users through a corpus rather than pinpointing exact information. We propose an exploratory search system based on hierarchical clusters and docume…