Apply video compression on KV cache to 10,000x less error at Q4 quantgithub.com·41 pts·polymorph1sm·7
Benchmarking the continuous improvement of language agents in deploymentarxiv.org·2 pts·polymorph1sm·0
Domain-Adaptive Object Detection via Uncertainty-Aware Distribution Alignmentgithub.com·2 pts·polymorph1sm·0
Intel 14nm and TSMC 7nm FinFET under Scanning Electron Microscopy [video]youtube.com·4 pts·polymorph1sm·0
Zhenhua Data leak: personal details of millions gathered by China companytheguardian.com·4 pts·polymorph1sm·0
Zero and DeepSpeed: system optimizations allow training models 100B parametersmicrosoft.com·1 pts·polymorph1sm·0