arxiv:2610.37062
Masato Fujitake
·
AI & ML interests
None yet
Recent Activity
authored a paper about 21 hours ago
EnergyCal: Taming Loss Spikes in Muon-Family Optimizers for Sparse MoE Training authored a paper about 21 hours ago
MESH: Memory-Efficient Sinkhorn Optimization for Mixture-of-Experts Training authored a paper about 21 hours ago
DTrOCR: Decoder-only Transformer for Optical Character Recognition