Metric Learning On Cub 200 2011

R@1

评测结果

各个模型在此基准测试上的表现结果

比较表格

模型名称	R@1
local-similarity-aware-deep-feature-embedding	58.3
integrating-language-guidance-into-vision	71.4
improved-embeddings-with-easy-positive	57.3
proxy-anchor-loss-for-deep-metric-learning	71.1
hard-negative-examples-are-hard-but-useful	57.7
proxynca-revisiting-and-revitalizing-proxy	69.0
the-group-loss-for-deep-metric-learning	65.5
attention-based-ensemble-for-deep-metric	60.6
metric-learning-with-horde-high-order	66.8
calibrated-neighborhood-aware-confidence	74.9
improved-embeddings-with-easy-positive	64.9
metric-learning-cross-entropy-vs-pairwise	69.2
learning-intra-batch-connections-for-deep	71.8
learning-semantic-proxies-from-visual-prompts	88.5
non-isotropy-regularization-for-proxy-based	70.5
center-contrastive-loss-for-metric-learning	73.45
towards-interpretable-deep-metric-learning	68.15
pads-policy-adapted-sampling-for-visual	67.3
unicom-universal-and-compact-representation	90.1
hardness-aware-deep-metric-learning	53.7
hyperbolic-vision-transformers-combining	85.6
diva-diverse-visual-feature-aggregation	69.2
s2sd-simultaneous-similarity-based-self	70.1
attributable-visual-similarity-learning	71.9
dissecting-the-impact-of-different-loss	63.8
it-takes-two-to-tango-mixup-for-deep-metric	71.4
sampling-matters-in-deep-embedding-learning	63.6
hard-aware-deeply-cascaded-embedding	60.7
mic-mining-interclass-characteristics-for	66.1
softtriple-loss-deep-metric-learning-without	65.4