Optimizing Probabilistic Models with Latent Variables

Accelerated Low Bit-Width Neural Network Training

Neural Network Training Trick Improves Model Generalization

Mathematical Foundations to Neural Network Optimization

Hard Drive Free ONNX IO

User Experience on LibTorch C++ API

Enable Parallelized Non-Autoregressive Decoding Process for Sequence to Sequence Tasks

Understand Autoregressive Models for Sequence to Sequence Tasks

Mathematical Foundations to Neural Network Pruning

Using Bootstrap Distribution to Approximate Sampling Distribution