Unverified50% confidenceFactExact time
DeepMind used a large-scale distributed training cluster for StarCraft II because a single machine could not complete training within an acceptable timeframe.
1
Sources
50%
Confidence
Long-term
Relevance
8/9/2026
First Seen
Sources
强化学习能打通无伤Boss战吗？PPO算法实战分析
Related Claims
UnverifiedDeepMind employed League Training in AlphaStar, maintaining an opponent pool of historical versions at different skill levels and styles to avoid overfitting to a single strategy.75% similarUnverifiedDeepSeek等国产模型成本突破得益于MoE稀疏架构、FlashAttention注意力加速技术和高效分布式训练调度策略70% similarUnverifiedThe chips DeepSeek is developing in-house are aimed at inference rather than training.70% similarVerifiedWhen developing Gemini, Google DeepMind adopted a multimodal training strategy built from scratch, enabling it to directly understand non-textual information such as temporal changes in video and tonal emotions in audio.70% similarUnverifiedDeepSpeed and Megatron-LM are mainstream large model training frameworks.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/714646API
curl https://kongchang.com/api/v1/knowledge/claims/714646MCP
get_claim(id=714646)