Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

LRA

Benchmarks

Task NameDataset NameSOTA ResultTrend
Long sequence classificationLRA (Long Range Arena) (test)
Average Accuracy87.09
92
Text ClassificationLRA Text (test)
Throughput (Steps/s)17.7
32
Long-range sequence modelingLRA 92 (test)
ListOps Accuracy63.14
26
ClassificationLRA Image N=1024 (test)
Accuracy87.3
23
Image ClassificationLRA reimplementation (test)
Running Time (h)2.19
9
RetrievalLRA Document Retrieval reimplementation (test)
Running Time (h)1.13
9
ListOpsLRA ListOps (test)
Running Time (h)0.6
9
Text ClassificationLRA Text Classification reimplementation (test)
Running Time (Hours)0.65
9
Text RetrievalLRA Text Retrieval @8000
Accuracy80.9
8
Document RetrievalLRA Retrieval
Accuracy82.3
8
Text ClassificationLRA Text
Accuracy68.4
8
Sequence ModelingLRA ListOps
Accuracy38.36
8
Long-Range RetrievalLRA Benchmark
ListOps Accuracy67.9
4
Showing 13 of 13 rows