Sehyun Choi's repositories
Chain-of-Hindsight-PyTorch
Unofficial implementation of Chain of Hindsight (https://arxiv.org/abs/2302.02676) using pytorch and huggingface Trainers.
An-Application-of-Deep-Reinforcement-Learning-to-Algorithmic-Trading
Experimental code supporting the results presented in the scientific research paper entitled "An Application of Deep Reinforcement Learning to Algorithmic Trading"
face_mask_inpaint
GAN-based project of removing facial masks from face images.
MAFIA-Explanation-NLI
Official code for MAFIA (MAsk-based Feature Interaction Attribution): Explain BERT model's decision in NLI task.
prompt-compression
This is an unofficial implementation of prompt-compression for my own research.
bittensor
Internet-scale Neural Networks
ColossalAI
Making large AI models cheaper, faster and more accessible
Diffusion-LM
Diffusion-LM
DSI-transformers
A huggingface transformers implementation of "Transformer Memory as a Differentiable Search Index"
gradio_serving
Simple example of using gradio for serving ML models.
HKUST-KSA-ML-Study
HKUST KSA 2020 ML study repository
HKUST_PhD_MPhil_thesis_Latex
The Hong Kong University of Science and Technology PhD/MPhil thesis latex template based on the latest official sample (http://pg.ust.hk/guides_n_forms/students/thesis_sample_page_phd.pdf)
jekyll-pygments-themes
CSS themes for Pygments syntax highlighter, ready to drop into Jekyll. Also includes a custom theme builder!
litgpt
Pretrain, finetune, deploy 20+ LLMs on your own data. Uses state-of-the-art techniques: flash attention, FSDP, 4-bit, LoRA, and more.
predicting-business-popularity
About Research project on predicting business popularity using GNNs in location-based social networks.
pretraining
Pretraining
RMT
Implementation of Recurrent Memory Transformer + Gating algorithms applied on long-context dialogue setting.
syncdoth.github.io
A beautiful, simple, clean, and responsive Jekyll theme for academics
tortoise-tts
A multi-voice TTS system trained with an emphasis on quality
transformers
🤗 Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX.