Bo-Yi Tsai's starred repositories
references-plus
References Plus
WSLAttachSwitch
Attach the WSL2 virtual machine to a Hyper-V virtual switch.
drag-select
jQuery based drag-select module, windows/mac like file selection module for web UI.
DragSelect
An easy JavaScript library for selecting and moving elements. With no dependencies. Drag-Select & Drag-And-Drop. – Examples:
wsl-kernel-modules
Build kernel modules for WSL2 distros
systemless-gapps
Systemless GApps converts specified NikGApps package into a module usable with Magisk or KernelSU
MagiskGapps
MagiskGApps is a modification of NikGApps that can be flashed in Magisk. It is built using a script that converts the NikGApps package into a MagiskGApps Package.
treble_experimentations
Notes about tinkering with Android Project Treble
TNP-pytorch
Official implementation of Transformer Neural Processes
pytorch_optimizer
optimizer & lr scheduler & loss function collections in PyTorch
sentence_representation_matching
句子匹配模型,包括无监督的SimCSE、ESimCSE、PromptBERT,和有监督的SBERT、CoSENT。
alt-anisette-server
Anisette Data as Service, like Sideloadly's, based on AltServer-Windows!
quarto-cli
Open-source scientific and technical publishing system built on Pandoc.
Keemap-Blender-Rig-ReTargeting-Addon
Blender Rig Retargeting Addon allows mapping motions of one rig to another. Works with ANY rig and allows user to map bones from one rig to another and save mapping files out to hard drive.
blender_mmd_uuunyaa_tools
mmd_uuunyaa_tools is a blender addon for adjust scenes, models and materials in concert with mmd_tools
BlendArMocap
realtime motion tracking in blender using mediapipe and rigify
MocapNET
We present MocapNET, a real-time method that estimates the 3D human pose directly in the popular Bio Vision Hierarchy (BVH) format, given estimations of the 2D body joints originating from monocular color images. Our contributions include: (a) A novel and compact 2D pose NSRM representation. (b) A human body orientation classifier and an ensemble of orientation-tuned neural networks that regress the 3D human pose by also allowing for the decomposition of the body to an upper and lower kinematic hierarchy. This permits the recovery of the human pose even in the case of significant occlusions. (c) An efficient Inverse Kinematics solver that refines the neural-network-based solution providing 3D human pose estimations that are consistent with the limb sizes of a target person (if known). All the above yield a 33% accuracy improvement on the Human 3.6 Million (H3.6M) dataset compared to the baseline method (MocapNET) while maintaining real-time performance
Obsidian-Table-Generator
A plugin for generate markdown table quickly like Typora.
fast-style-transfer
TensorFlow CNN for fast style transfer ⚡🖥🎨🖼