Similar Items: Parallel-in-Time Training of Recurrent Neural Networks for Dynamical Systems Reconstruction
- AutoSP: Unlocking Long-Context LLM Training Via Compiler-Based Sequence Parallelism
- Piper: Efficient Large-Scale MoE Training via Resource Modeling and Pipelined Hybrid Parallelism
- Towards Compute-Aware In-Switch Computing for LLMs Tensor-Parallelism on Multi-GPU Systems
- Regulating Branch Parallelism in LLM Serving
- ResiHP: Taming LLM Training Failures with Dynamic Hybrid
- ShardTensor: Domain Parallelism for Scientific Machine Learning