Open-Source Activities

CANVAS

CANVAS: Commonsense-Aware Navigation System for Intuitive Human-Robot Interaction

HerO

HerO at AVeriTeC: The Herd of Open Large Language Models for Verifying Real-World Claims

EnCLAP

EnCLAP: Combining Neural Audio Codec and Audio-Text Joint Embedding for Automated Audio Captioning

PhaseAug

PhaseAug: A Differentiable Augmentation for Speech Synthesis to Simulate One-to-Many Mapping

Assem-VC

Assem-VC: Realistic Voice Conversion by Assembling Modern Speech Synthesis Techniques

NU-Wave

NU-Wave: A Diffusion Probabilistic Model for Neural Audio Upsampling

NU-Wave 2

NU-Wave 2: A General Neural Audio Upsampling Model for Various Sampling Rates

Cotatron

Cotatron: Transcription-Guided Speech Encoder for Any-to-Many Voice Conversion without Parallel Data

pNLP-Mixer

pNLP-Mixer: an Efficient all-MLP Architecture for Language

First successful open-source implementation of pNLP-Mixer.

HifiFace

HifiFace: 3D Shape and Semantic Prior Guided High Fidelity Face Swapping

First successful open-source implementation of HifiFace.

UnivNet

UnivNet: A Neural Vocoder with Multi-Resolution Spectrogram Discriminators for High-Fidelity Waveform Generation

First successful open-source implementation of UnivNet.

WaveGrad2

WaveGrad 2: Iterative Refinement for Text-to-Speech Synthesis

First successful open-source implementation of WaveGrad2.

FaceShifter

FaceShifter: Towards High Fidelity And Occlusion Aware Face Swapping

First successful open-source implementation of FaceShifter.

Reformer-pytorch

Reformer: The Efficient Transformer

Implementation of Reformer in PyTorch.

MelGAN

MelGAN: Generative Adversarial Networks for Conditional Waveform Synthesis

Implementation of MelGAN vocoder (compatible with NVIDIA/tacotron2)

MelNet

MelNet: A Generative Model for Audio in the Frequency Domain

Implementation of MelNet. Work done with Deepest AI (SNU Deep Learning Society).

RandWireNN

Exploring Randomly Wired Neural Networks for Image Recognition

Implementation of RandWireNN.

VoiceFilter

VoiceFilter: Targeted Voice Separation by Speaker-Conditioned Spectrogram Masking

First successful open-source implementation of Google's VoiceFilter

KoTDG

Korean Text Data Generator for OCR tasks.

Data-Science-for-COVID-19

COVID-19 Korea Dataset with patient routes and visualizer. Co-led the collaboration project.

Alias-Free-Torch

Simple torch.nn.module implementation of Alias-Free-GAN style filter and resample.