Tensor and Sequence Parallelism
PulseAugur coverage of Tensor and Sequence Parallelism — every cluster mentioning Tensor and Sequence Parallelism across labs, papers, and developer communities, ranked by signal.
1 day(s) with sentiment data
-
LLM framework jointly evolves routing optimization heuristics
Researchers have developed a novel framework called LLM-Driven Heuristic Components Joint Generation (LLM-HCJG) to address the limitations of expert knowledge in heuristic design for combinatorial optimization. This pop…
-
LINC architecture improves routing solver performance on CVRPTW, TSP, and CVRP
Researchers have introduced LINC (Local Inference via Normed Comparison), a novel architecture for constructive neural routing solvers. LINC explicitly computes one-step consequences like travel and capacity changes, de…
-
Zyphra's TSP strategy boosts LLM training throughput by 2.6x
Zyphra has developed a new technique called Tensor and Sequence Parallelism (TSP) designed to optimize the training and inference of large transformer models. This hardware-aware strategy combines aspects of Tensor Para…
-
New TSP strategy folds tensor and sequence parallelism for memory-efficient training
Researchers have introduced a new parallel execution strategy called Tensor and Sequence Parallelism (TSP) designed to enhance memory efficiency during the training and inference of Transformer models. TSP combines tens…