Yang Song
Research Scientist at OpenAI
Palo Alto, California, United States
Summary
🤩
RockstarYang Song is a research scientist with 11 years of experience focused on machine learning and generative modeling, currently based in Palo Alto and affiliated with OpenAI. He has hands-on expertise implementing and refining score-based generative models, notably contributing engineering work to the official Score SDE repository for the ICLR 2021 oral paper. Yang’s contributions span training and sampling configuration, dataset-specific model tuning, and practical debugging and dependency upgrades—skills that bridge research ideas to reproducible code. He combines a research mindset with rigorous engineering practice, enabling reliable experiments at scale. An interesting detail: his public commits reveal a penchant for optimizing sampler predictor–corrector settings, highlighting a deep practical interest in the dynamics of stochastic differential equation–based generation.
11 years of coding experience
Github Skills (8)
diffusion-models10
generative-model10
machine-learning10
jax10
flax10
python9
tensorflow8
pytorch3
Programming languages (7)
Github contributions (5)
yang-song/score_sde
Feb 2021 - Aug 2021
Official code for Score-Based Generative Modeling through Stochastic Differential Equations (ICLR 2021, Oral)
Role in this project:
ML Engineer
ML EngineerContributions:24 commits, 17 pushes, 18 comments in 6 months
Contributions summary:Yang primarily focused on modifying and configuring files related to training and sampling for the Score-Based Generative Modeling project. Their commits demonstrate changes to model configurations for different datasets (FFHQ, CelebAHQ), and adjustments to the sampling process including predictor and corrector settings. They also addressed bugs related to model loading and upgraded dependencies, indicating active involvement in the model's implementation and optimization.
dynamical-systemsflaxmixture-of-expertsscore-based-generative-modelingdifferential
yang-song/score_sde_pytorch
Feb 2021 - Nov 2022
PyTorch implementation for Score-Based Generative Modeling through Stochastic Differential Equations (ICLR 2021, Oral)
Contributions:7 commits, 1 PR, 3 pushes in 1 year 8 months
pytorchstochastic-differential-equationsinverse-problemsgenerative-modelsscore-matching
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.