1 个仓库
Awesome GitHub RepositoriesModel Component Colocation
Techniques for sharing memory by colocating model roles and components on the same hardware.
Distinct from Training Memory Optimizers: Focuses on sharing memory via role-swapping/sleep-mode rather than general training algorithms like gradient checkpointing
Explore 1 awesome GitHub repository matching data & databases · Model Component Colocation. Refine with filters or upvote what's useful.
Awesome Model Component Colocation GitHub Repositories
用 AI 发现最棒的仓库。我们将通过 AI 为您搜索最匹配的仓库。
- openrlhf/openrlhf
OpenRLHF/OpenRLHF
9,675在 GitHub 上查看OpenRLHF is a training framework and alignment library designed for reinforcement learning from human feedback across distributed GPU clusters. It provides tools for aligning large language models and multimodal vision-language models using algorithms such as PPO, GRPO, and DPO. The framework distinguishes itself through a distributed inference engine that overlaps sample rollout with training to increase throughput. It supports scaling to models exceeding 70 billion parameters via parameter sharding and handles long-context sequences through ring-attention sequence parallelism. The project
Optimizes memory on small clusters by colocating model components and sharing resources via sleep-mode.
Pythonlarge-language-modelsopenai-o1proximal-policy-optimization
