ShangwuYao/Pose-guided-action-recognition
Use human pose information to help action recognition, explored with attention-pooling method, C3D method and two-stream architecture, implemented in pytorch.
Machine learning and deep learning
Use human pose information to help action recognition, explored with attention-pooling method, C3D method and two-stream architecture, implemented in pytorch.
Clspv is a prototype compiler for a subset of OpenCL C to Vulkan compute shaders
Flask + PostgreSQL + Docker + Travis-CI + Pytest
Using group lasso regularization in LSTM to create ISS, achieved 10x speedup
part 1~3
Convert ImageNet dataset to Tf-Record format
CMU MITS capstone project, assessing robustness of deep learning models against attacks
scikit-learn: machine learning in Python
support material for github issues
Implementation of fast exact k-means algorithms
Solving MontezumaRevenge with deep reinforcement learning
p1
Sparse LSTMs -- Learning Connections and Hidden Sizes (TensorFlow)
Preparing for serious Kaggle competitions!