Unverified50% confidenceFactExact time
The NanoGPT Speedrun concept originated from Andrej Karpathy's nanoGPT project.
1
Sources
50%
Confidence
Long-term
Relevance
9/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn the NanoGPT Speedrun, Muon is typically applied only to the matrix weights of hidden layers, while the embedding layer and output layer retain AdamW.64% similarUnverifiedNanoStats 可实时显示 CPU 使用率、内存占用以及网络上行与下行速度58% similarUnverifiedMoS-TTS-Nano可能采用了非自回归架构来平衡质量与推理速度56% similarUnverifiedFastSpeech（2019）采用非自回归结构实现并行生成，将合成速度提升数十倍56% similarUnverifiedIntel PIN、DynamoRIO等主流二进制插桩工具通过动态二进制翻译（DBT）实现细粒度追踪54% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/910671API
curl https://kongchang.com/api/v1/knowledge/claims/910671MCP
get_claim(id=910671)