Skip to content

Projects

8 open-source systems and research software artifacts, with GitHub impact shown where available.

ServerlessLLM

Checkpoint-aware serverless LLM serving for fast, cost-efficient custom model deployment.

684
machine-learning-systems

MegBA

Distributed GPU bundle-adjustment library for large-scale 3D reconstruction workloads.

491
machine-learning-systems

TorchOpt

PyTorch library for differentiable optimization, meta-learning, and implicit or zero-order gradients.

630
machine-learning-systems

Quiver

Low-latency GPU graph-learning runtime for scaling PyG workloads across machines.

304
machine-learning-systems

KungFu

Adaptive distributed training runtime with monitoring and control APIs for large GPU clusters.

295
machine-learning-systems

TensorLayer

Deep learning and reinforcement learning library with reusable layers, models, and training utilities.

7.4K
machine-learning-systems

HyperPose

Real-time pose-estimation framework with high-level APIs and optimized CPU/GPU execution.

1.3K
machine-learning-systems

RLzoo

Reinforcement-learning model zoo with ready-to-run algorithms, environments, and training utilities.

641
machine-learning-systems