Top 5% on Kaggle leaderboard using fast.ai library and resnet50 along with transfer learning.
-
Updated
Aug 24, 2018 - Jupyter Notebook
Top 5% on Kaggle leaderboard using fast.ai library and resnet50 along with transfer learning.
使用TensorFlow自己搭建一些经典的CNN模型,并使用统一的数据来测试效果。
Detecting Action performed in a video using resnet34 for spatial and temporal stream
MNIST classification with deeper CNN models
Classifies static images as wearing a face mask or not.
Image classifier based on Pytorch
Few-shot learning experiments mostly on speaker recognition.
CNN to recognize which of my cats appears in a photo.
This is an implementation of ResNet using keras.
Deep CNN models ResNet34 and VGG16 have been trained and tested for image classification task using MNIST and CIFAR dataset (part of mini-project from the Deep learning and computer vision module)
Explainable Speaker Recognition
AI based image classification inspired MobileNet V2 architecture by implementing changes in base architecture and details about using it as a quick response model (proposition) for rapid application as well as comparing it with other models for the same application.
Knocking occurs when fuel burns unevenly in your engine. When everything is going as it should, and the cylinders have the correct mix of air and fuel, the mixture burns in a controlled, progressive manner. After each cylinder's air/fuel mixture burns, it should create a small “shock wave” in your engine. This project is a knocking prediction app.
This repository contains Final project of CSE428 Brac University
Classifying waste types using transfer learning, with scripts for training, evaluation, and predictions
U-Net segmentation algorithm with options of pretrained resnet34 and resnet50 encoders. All of the project dockerized with gpu suppport on anaconda environment with multiple loss support..
Add a description, image, and links to the resnet34 topic page so that developers can more easily learn about it.
To associate your repository with the resnet34 topic, visit your repo's landing page and select "manage topics."