Papers › ResNet50_on_Cifar_100_Without_Transfer_Learning

ResNet50_on_Cifar_100_Without_Transfer_Learning

3 Aug 2020Sorry, no such thing happened 2020 8archive 2025-07-28

Batuhan Bayraktar

I am a young person who is curious about image classification. I have an average knowledge and an average computer. I thought the cifar 100 data for image classification was challenging for me. It was a data containing 100 classes from 32x32 and 100 images. I chose resnet as the model due to the low number of data and gradient vanishing problem. I worked with google colab because my computer is not enough (thank you Google) Since I used the free version, I could only run Resnet50. I tried to stick to the original Resnet article. But I made it myself in changes. I have tried many hyper parameters. I found the parameters that gave the best results as soon as possible as fast as I could. I'm a young man who likes to push his luck. that is all

PaperPDFCode

Code

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

General ClassificationImage Classificationimage-classification

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Image Classification CIFAR-100 ResNet50 Without Transfer Learning Percentage correct 67.060 #187 of 211 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

1x1 ConvolutionAMAverage PoolingBatch NormalizationBottleneck Residual BlockConvolutionGlobal Average PoolingKaiming InitializationMax PoolingReLUResidual BlockResidual Connection

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections