Accelerating a Triton Fused Kernel for W4A16 Quantized Inference with SplitK work decomposition.
Adnan HoqueLess WrightJamie YangMudhakar SrivatsaRaghu K. GantiPublished in: CoRR (2024)
Keyphrases
- data fusion
- kernel methods
- probabilistic inference
- bayesian inference
- bayesian networks
- kernel function
- feature space
- probabilistic model
- information fusion
- kernel regression
- fusion method
- decomposition method
- bayesian hierarchical model
- kernel space
- image decomposition
- decomposition methods
- reproducing kernel hilbert space
- probabilistic reasoning
- kernel learning
- kernel machines
- inference process
- decomposition algorithm
- wavelet packet
- inference engine
- random fields
- mutual subspace method
- knowledge base