Codesota · Models · ProphetNet-XUSTC / Microsoft Research Asia1 results · 1 benchmarks
Model card

ProphetNet-X.

USTC / Microsoft Research Asiacode-lmTransformer encoder-decoderMIT

Large-scale pre-training model for code generation (ProphetNet-Code variant). From arXiv 2104.08006.

§ 01 · Benchmarks

Every benchmark ProphetNet-X has a recorded score for.

#BenchmarkArea · TaskMetricValueRankDateSource
01codesearchnet---javaComputer Vision · Optical Character Recognitionsmoothed-bleu-419.4%#5/14—source ↗
Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.
§ 02 · Strengths by area

Where ProphetNet-X actually performs.

Computer Vision
1
benchmark
avg rank #5.0
§ 05 · Sources & freshness

Where these numbers come from.

codexglue-leaderboard
1
result
1 of 1 rows marked verified.