All Projects → digitalbrain79 → resnet-ensemble

digitalbrain79 / resnet-ensemble

Licence: other
Ensemble code for Resnet in Tensorflow slim

Programming Languages

python
139335 projects - #7 most used programming language
shell
77523 projects

Projects that are alternatives of or similar to resnet-ensemble

AESRC2020
a deep accent recognition network
Stars: ✭ 35 (+150%)
Mutual labels:  resnet
Machine-learning-toolkits-with-python
Machine learning toolkits with Python
Stars: ✭ 31 (+121.43%)
Mutual labels:  ensemble
neural-dream
PyTorch implementation of DeepDream algorithm
Stars: ✭ 110 (+685.71%)
Mutual labels:  resnet
Water-classifier-fastai
Deploy your Flask web app classifier on Heroku which is written using fastai library.
Stars: ✭ 37 (+164.29%)
Mutual labels:  resnet
TensorMONK
A collection of deep learning models (PyTorch implemtation)
Stars: ✭ 21 (+50%)
Mutual labels:  resnet
BMW-IntelOpenVINO-Detection-Inference-API
This is a repository for a No-Code object detection inference API using the OpenVINO. It's supported on both Windows and Linux Operating systems.
Stars: ✭ 66 (+371.43%)
Mutual labels:  resnet
Keras-MultiClass-Image-Classification
Multiclass image classification using Convolutional Neural Network
Stars: ✭ 48 (+242.86%)
Mutual labels:  resnet
TrackNet-Badminton-Tracking-tensorflow2
TrackNet for badminton tracking using tensorflow2
Stars: ✭ 37 (+164.29%)
Mutual labels:  resnet
simpleAICV-pytorch-ImageNet-COCO-training
SimpleAICV:pytorch training example on ImageNet(ILSVRC2012)/COCO2017/VOC2007+2012 datasets.Include ResNet/DarkNet/RetinaNet/FCOS/CenterNet/TTFNet/YOLOv3/YOLOv4/YOLOv5/YOLOX.
Stars: ✭ 276 (+1871.43%)
Mutual labels:  resnet
ModelZoo.pytorch
Hands on Imagenet training. Unofficial ModelZoo project on Pytorch. MobileNetV3 Top1 75.64🌟 GhostNet1.3x 75.78🌟
Stars: ✭ 42 (+200%)
Mutual labels:  resnet
Deep-Vessel
kgpml.github.io/deep-vessel/
Stars: ✭ 52 (+271.43%)
Mutual labels:  ensemble
MVCNN-PyTorch
No description or website provided.
Stars: ✭ 79 (+464.29%)
Mutual labels:  resnet
subsemble
subsemble R package for ensemble learning on subsets of data
Stars: ✭ 40 (+185.71%)
Mutual labels:  ensemble
pb-gcn
Code for the BMVC paper (http://bmvc2018.org/contents/papers/1003.pdf)
Stars: ✭ 32 (+128.57%)
Mutual labels:  resnet
Deep-Compression.Pytorch
Unofficial Pytorch implementation of Deep Compression in CIFAR10
Stars: ✭ 29 (+107.14%)
Mutual labels:  resnet
bagging pu
Simple sklearn based python implementation of Positive-Unlabeled (PU) classification using bagging based ensembles
Stars: ✭ 73 (+421.43%)
Mutual labels:  ensemble
keras-stochastic-weight-averaging
Keras callback function for stochastic weight averaging
Stars: ✭ 53 (+278.57%)
Mutual labels:  ensemble
neuralBlack
A Multi-Class Brain Tumor Classifier using Convolutional Neural Network with 99% Accuracy achieved by applying the method of Transfer Learning using Python and Pytorch Deep Learning Framework
Stars: ✭ 36 (+157.14%)
Mutual labels:  resnet
resnet-cifar10
ResNet for Cifar10
Stars: ✭ 21 (+50%)
Mutual labels:  resnet
Look4Face
Demo of Face Recognition web service
Stars: ✭ 23 (+64.29%)
Mutual labels:  resnet

Resnet Ensemble

tensorflow/models/slim 에서 evaluation 하는데 필요한 파일만 수정해서 올렸습니다. Ensemble은 각 model에서 나온 output을 평균한 후 softmax에 넣는 방법으로 했습니다.

Dataset

Food validation data

Checkpoints Files

Resnet 50 checkpoints files

Evaluation

DATASET_DIR=data
CHECKPOINT_DIR=checkpoints/resnet_50_1_100000.ckpt,checkpoints/resnet_50_2_100000.ckpt,checkpoints/resnet_50_3_100000.ckpt
python eval_image_classifier_ensemble.py \
	--alsologtostderr \
	--checkpoint_path=${CHECKPOINT_DIR} \
	--dataset_dir=${DATASET_DIR} \
	--dataset_name=food \
	--dataset_split_name=validation \
	--model_name=resnet_v2_50

Result

Model Top 1 Accuracy Top 5 Accuracy
Resnet_50_1 0.676 0.885
Resnet_50_2 0.674 0.877
Resnet_50_3 0.675 0.880
Ensemble 0.738 0.914
위의 결과와 같이 ensemble을 하니 top 1 accuracy가 0.6정도 증가하고 top 5 accuracy가 0.3정도 증가했습니다.
Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].