How to Scale Your Model：大模型扩展完整指南 on Pengcheng's Bloghttps://pengchengneo.github.io/posts/scaling-book-cn/Recent content in How to Scale Your Model：大模型扩展完整指南 on Pengcheng's BlogHugo -- gohugo.iozh-CN© 2026 PengchengWed, 29 Apr 2026 00:00:00 +0000Introduction（引言）https://pengchengneo.github.io/posts/scaling-book-cn/part0_introduction/Wed, 29 Apr 2026 00:00:00 +0000https://pengchengneo.github.io/posts/scaling-book-cn/part0_introduction/Intro to Rooflines（Roofline 模型入门）https://pengchengneo.github.io/posts/scaling-book-cn/part1_roofline/Wed, 29 Apr 2026 00:00:00 +0000https://pengchengneo.github.io/posts/scaling-book-cn/part1_roofline/All About TPUs（TPU 详解）https://pengchengneo.github.io/posts/scaling-book-cn/part2_tpus/Wed, 29 Apr 2026 00:00:00 +0000https://pengchengneo.github.io/posts/scaling-book-cn/part2_tpus/Sharded Matmuls（分片矩阵乘法）https://pengchengneo.github.io/posts/scaling-book-cn/part3_sharding/Wed, 29 Apr 2026 00:00:00 +0000https://pengchengneo.github.io/posts/scaling-book-cn/part3_sharding/Transformershttps://pengchengneo.github.io/posts/scaling-book-cn/part4_transformers/Wed, 29 Apr 2026 00:00:00 +0000https://pengchengneo.github.io/posts/scaling-book-cn/part4_transformers/Training（训练）https://pengchengneo.github.io/posts/scaling-book-cn/part5_training/Wed, 29 Apr 2026 00:00:00 +0000https://pengchengneo.github.io/posts/scaling-book-cn/part5_training/Training LLaMA（训练 LLaMA）https://pengchengneo.github.io/posts/scaling-book-cn/part6_applied_training/Wed, 29 Apr 2026 00:00:00 +0000https://pengchengneo.github.io/posts/scaling-book-cn/part6_applied_training/Inference（推理）https://pengchengneo.github.io/posts/scaling-book-cn/part7_inference/Wed, 29 Apr 2026 00:00:00 +0000https://pengchengneo.github.io/posts/scaling-book-cn/part7_inference/Serving LLaMA（部署 LLaMA）https://pengchengneo.github.io/posts/scaling-book-cn/part8_applied_inference/Wed, 29 Apr 2026 00:00:00 +0000https://pengchengneo.github.io/posts/scaling-book-cn/part8_applied_inference/Profiling（性能分析）https://pengchengneo.github.io/posts/scaling-book-cn/part9_profiling/Wed, 29 Apr 2026 00:00:00 +0000https://pengchengneo.github.io/posts/scaling-book-cn/part9_profiling/All About JAX（JAX 详解）https://pengchengneo.github.io/posts/scaling-book-cn/part10_jax/Wed, 29 Apr 2026 00:00:00 +0000https://pengchengneo.github.io/posts/scaling-book-cn/part10_jax/Conclusions（总结）https://pengchengneo.github.io/posts/scaling-book-cn/part11_conclusion/Wed, 29 Apr 2026 00:00:00 +0000https://pengchengneo.github.io/posts/scaling-book-cn/part11_conclusion/GPUshttps://pengchengneo.github.io/posts/scaling-book-cn/part12_gpus/Wed, 29 Apr 2026 00:00:00 +0000https://pengchengneo.github.io/posts/scaling-book-cn/part12_gpus/