Leolewis5/Machine_Learning

Record for personal machine learning note

Machine_Learning

This is my personal learning note, recorded of deep learning study. Also, record of kaggle compitition result and code.

Github sometimes has issue for opening jupyter, therefore, items would be opened through: https://nbviewer.jupyter.org/

Requirements

• Anaconda3(numpy/pandas/matplotlib)
• tensorflow-gpu 1.13.1
• keras 2.2.4
• nvidia CUDA 10.0
• cudnn v7.5.0
• pyroch 1.2.0

Python data handle module

• Pandas
• Matplotlib

Deep Learning with Keras

• 0. Bisic regression
• 1. MNIST DNN
• 2. MNIST CNN
• 3. VGG16_model construction
• 4. Transfer learning-base on VGG16
• 5. Small volume data training-Data Augumentation
• 6. Visualization CNN output-VGG16
• 7. AutoEncoder-DNN dimension reduction
• 8. Style Transfer-based on VGG19
• 9. Batch Normalization
• 10. GAN - MNIST
• 11. DCGAN - Animation Face
• 12. cDCGAN-MNIST
• 13. Stochastic Weight Averaging (SWA)
• 14. Squeeze-and-Excitation Networks (SE-NET)
• 15. Convolutional Block Attention Module (CBAM)
• 16. UNET

Deep Learning with Pytorch

• 1. CNN_MNIST
• 2. AnimiFace_RALSAGIN
• 3. BIG_GAN_deep_128_pretrain

Kaggle

• 1. Titanic: Machine Learning from Disaster-DNN/ tabular data analysis
• 2. Generative Dog Images-RALSGAN
• 3. SIIM_ACR Pneumothorax Segmentation-Medical image segmentation
• 4. Severstal Steel Defect Detection-Degect image segmentation and classification
• 5. RSNA Intracranial Hemorrhage Detection