FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3206343543> ?p ?o ?g. }

• W3206343543 abstract "There is often variation in the shape and size of input data used for deep learning. In many cases, such data can be represented using tensors with non-uniform shapes, or ragged tensors. Due to limited and non-portable support for efficient execution on ragged tensors, current deep learning frameworks generally use techniques such as padding and masking to make the data shapes uniform and then offload the computations to optimized kernels for dense tensor algebra. Such techniques can, however, lead to a lot of wasted computation and therefore, a loss in performance. This paper presents CoRa, a tensor compiler that allows users to easily generate efficient code for ragged tensor operators targeting a wide range of CPUs and GPUs. Evaluating CoRa on a variety of operators on ragged tensors as well as on an encoder layer of the transformer model, we find that CoRa (i)performs competitively with hand-optimized implementations of the operators and the transformer encoder and (ii) achieves, over PyTorch, a 1.6X geomean speedup for the encoder on an Nvidia GPU and a 1.86X geomean speedup for the multi-head attention module used in transformers on an ARM CPU." @default.
• W3206343543 created "2021-10-25" @default.
• W3206343543 creator A5010837144 @default.
• W3206343543 creator A5014824446 @default.
• W3206343543 creator A5063105034 @default.
• W3206343543 creator A5076667915 @default.
• W3206343543 date "2021-10-19" @default.
• W3206343543 modified "2023-09-27" @default.
• W3206343543 title "The CoRa Tensor Compiler: Compilation for Ragged Tensors with Minimal Padding" @default.
• W3206343543 cites W131533222 @default.
• W3206343543 cites W1480909796 @default.
• W3206343543 cites W1667652561 @default.
• W3206343543 cites W2055312318 @default.
• W3206343543 cites W2099021415 @default.
• W3206343543 cites W2194775991 @default.
• W3206343543 cites W2245094585 @default.
• W3206343543 cites W2471164860 @default.
• W3206343543 cites W2519091744 @default.
• W3206343543 cites W2549815592 @default.
• W3206343543 cites W2590246587 @default.
• W3206343543 cites W2606964149 @default.
• W3206343543 cites W2626778328 @default.
• W3206343543 cites W2752134738 @default.
• W3206343543 cites W2786320458 @default.
• W3206343543 cites W2804500013 @default.
• W3206343543 cites W2891555348 @default.
• W3206343543 cites W2891815651 @default.
• W3206343543 cites W2898106867 @default.
• W3206343543 cites W2905135312 @default.
• W3206343543 cites W2913790721 @default.
• W3206343543 cites W2925102126 @default.
• W3206343543 cites W2949251082 @default.
• W3206343543 cites W2950813464 @default.
• W3206343543 cites W2953109491 @default.
• W3206343543 cites W2953384591 @default.
• W3206343543 cites W2961619211 @default.
• W3206343543 cites W2963323070 @default.
• W3206343543 cites W2963341956 @default.
• W3206343543 cites W2963355447 @default.
• W3206343543 cites W2963846996 @default.
• W3206343543 cites W2970971581 @default.
• W3206343543 cites W2978670439 @default.
• W3206343543 cites W2979044977 @default.
• W3206343543 cites W2984082758 @default.
• W3206343543 cites W3034107927 @default.
• W3206343543 cites W3036692157 @default.
• W3206343543 cites W3039314686 @default.
• W3206343543 cites W3090487264 @default.
• W3206343543 cites W3128966618 @default.
• W3206343543 cites W3130660608 @default.
• W3206343543 cites W3136585513 @default.
• W3206343543 cites W3136805225 @default.
• W3206343543 cites W3136964960 @default.
• W3206343543 cites W3177452048 @default.
• W3206343543 cites W3196320218 @default.
• W3206343543 cites W3197289565 @default.
• W3206343543 cites W3207857196 @default.
• W3206343543 hasPublicationYear "2021" @default.
• W3206343543 type Work @default.
• W3206343543 sameAs 3206343543 @default.
• W3206343543 citedByCount "0" @default.
• W3206343543 crossrefType "posted-content" @default.
• W3206343543 hasAuthorship W3206343543A5010837144 @default.
• W3206343543 hasAuthorship W3206343543A5014824446 @default.
• W3206343543 hasAuthorship W3206343543A5063105034 @default.
• W3206343543 hasAuthorship W3206343543A5076667915 @default.
• W3206343543 hasConcept C100376341 @default.
• W3206343543 hasConcept C111919701 @default.
• W3206343543 hasConcept C11413529 @default.
• W3206343543 hasConcept C118505674 @default.
• W3206343543 hasConcept C121332964 @default.
• W3206343543 hasConcept C136119220 @default.
• W3206343543 hasConcept C155281189 @default.
• W3206343543 hasConcept C165435473 @default.
• W3206343543 hasConcept C165801399 @default.
• W3206343543 hasConcept C1680195 @default.
• W3206343543 hasConcept C169590947 @default.
• W3206343543 hasConcept C173608175 @default.
• W3206343543 hasConcept C199360897 @default.
• W3206343543 hasConcept C202444582 @default.
• W3206343543 hasConcept C2524010 @default.
• W3206343543 hasConcept C33923547 @default.
• W3206343543 hasConcept C38652104 @default.
• W3206343543 hasConcept C41008148 @default.
• W3206343543 hasConcept C45374587 @default.
• W3206343543 hasConcept C459310 @default.
• W3206343543 hasConcept C62520636 @default.
• W3206343543 hasConcept C66322947 @default.
• W3206343543 hasConcept C68339613 @default.
• W3206343543 hasConcept C81999800 @default.
• W3206343543 hasConceptScore W3206343543C100376341 @default.
• W3206343543 hasConceptScore W3206343543C111919701 @default.
• W3206343543 hasConceptScore W3206343543C11413529 @default.
• W3206343543 hasConceptScore W3206343543C118505674 @default.
• W3206343543 hasConceptScore W3206343543C121332964 @default.
• W3206343543 hasConceptScore W3206343543C136119220 @default.
• W3206343543 hasConceptScore W3206343543C155281189 @default.
• W3206343543 hasConceptScore W3206343543C165435473 @default.
• W3206343543 hasConceptScore W3206343543C165801399 @default.
• W3206343543 hasConceptScore W3206343543C1680195 @default.

• next