res_d759d80cAn ensemble of these residual nets achieves 3.57% error on the ImageNettest set.
arXiv 1512.03385 · cs.CV
Read on arXiv · arXiv non-exclusive license
Claims were extracted automatically by reducer-1 on 2026-11-12. The paper's license does not allow republishing it, so long claims are shortened to a quotation. Read the full text on arXiv.
res_d759d80cAn ensemble of these residual nets achieves 3.57% error on the ImageNettest set.
res_c1f23784Solely due to our extremely deep representations, we obtain a 28% relative improvement on the COCO object detection dataset.
res_027ebee5Unexpectedly, such degradation is not caused by overfitting , and adding more layers to a suitably deep model leads to higher training error, as reported in [11, 42] and thoroughly … (excerpt)
clm_fe8c998cWe present comprehensive experiments on ImageNet [36] to show the degradation problem and evaluate our method.
res_06c2da64We show that: 1) Our extremely deep residual nets are easy to optimize, but the counterpart “plain” nets (that simply stack layers) exhibit higher training error when the depth increases; … (excerpt)
res_33738518Our ensemble has 3.57% top-5 error on the ImageNet test set, and won the 1st place in the ILSVRC 2015 classification competition .
res_00f66dd2The results in Table 2 show that the deeper 34-layer plain net has higher validation error than the shallower 18-layer plain net.
res_70d31aee34-layer plain net has higher training error throughout the whole training procedure, even though the solution space of the 18-layer plain network is a subspace of that of the 34-layer … (excerpt)
clm_9346a7c0We argue that this optimization difficulty is unlikely to be caused by vanishing gradients.
res_00908340In fact, the 34-layer plain net is still able to achieve competitive accuracy (Table 3), suggesting that the solver works to some extent.
res_beb44ca7First, the situation is reversed with residual learning – the 34-layer ResNet is better than the 18-layer ResNet (by 2.8%).
res_0f3d3feeMore importantly, the 34-layer ResNet exhibits considerably lower training error and is generalizable to the validation data.
clm_0c33a7d3We argue that this is because the zero-padded dimensions in A indeed have no residual learning.
res_d62afd15Remarkably, although the depth is significantly increased, the 152-layer ResNet (11.3 billion FLOPs) still has lower complexity than VGG-16/19 nets (15.3/19.6 billion FLOPs).
res_43a9f1a4Our baseline 34-layer ResNets have achieved very competitive accuracy.
res_7e1194bdThis single-model result outperforms all previous ensemble results (Table 5).
res_c630da1eIt has fewer parameters than other deep and thin 5With an initial learning rate of 0.1, it starts converging ( <90% error) after several epochs, but still reaches similar accuracy.
res_f3fe2ba3The error of plain-110 is higher than 60% and not displayed.
res_3bef384aOur method shows no optimization difficulty , and this 103-layer network is able to achieve training error <0.1% (Fig. 6, right).
res_43aed781Its test error is still fairly good (7.93%, Table 6).
res_73b8f902Object detection mAP (%) on the PASCAL VOC 2007/2012 test sets using baseline Faster R-CNN.
res_7a376285See also Table 9 for better results. have similar training error.
clm_f1d76064We argue that this is because of overfitting.
res_cf2b5755Here we are interested in the improvements of replacing VGG-16 [41] with ResNet-101.
res_24dca08dMost remarkably, on the challenging COCO dataset we obtain a 6.0% increase in COCO’s standard metric (mAP@[.5, .95]), which is a 28% relative improvement.
res_8eb755bbResNet-101 has a 6% increase of mAP@[.5, .95] over VGG-16, which is a 28% relative improvement, solely contributed by the features learned by the better network.
clm_754f6c20For completeness, we report the improvements made for the competitions.
res_91fce645Box refinement improves mAP by about 2 points (Table 9).
res_4722bc81Global context improves mAP@.5 by about 1 point (Table 9).
res_1f74fc32Multi-scale testing improves the mAP by over 2 points (Table 9).
res_744ed5f5Our detection system is Faster R-CNN [32] with the improvements in Table 9, using ResNet-101. we achieve 85.6% mAP on PASCAL VOC 2007 (Table 10) and 83.8% on PASCAL VOC … (excerpt)
res_23b3f42358.8% mAP and our ensemble of 3 models has 62.1% mAP on the DET test set (Table 12).This result won the 1st place in the ImageNet detection task in ILSVRC … (excerpt)
res_bed0d1e2Using an ensemble of networks for both classification and localization, we achieve a top-5 localization error of 9.0% on the test set.
res_d7de4b17This number significantly outperforms the ILSVRC 14 results (Table 14), showing a 64% relative reduction of error.
Record your result in the Mearc app on your own machine, then use Publish this result with this paper's id (1512.03385) and the claim id shown next to the claim. Mearc creates a file that holds only what you choose to publish. Add it to the public claim data by pull request; a maintainer other than you reviews it before it counts.