Tag: 混合专家架构 (2 articles)

NVIDIA NeMo AutoModel seamlessly plugs into the HuggingFace ecosystem, boosting MoE fine-tuning throughput by 3.4x-3.7x and cutting VRAM usage by 30% with a single import line change.

Z.ai releases the 753B-parameter open-weights GLM-5.2, topping key benchmarks while consuming excessive tokens, signaling a new era of brute-force open-source AI.