bojieli/ai-infra-book

★8,137Python🇨🇳 Greater ChinaApache-2.0⑂ 661Issue 6Updated Oct 1, 2026

《深入理解 AI Infra:量化分析与系统设计》(李博杰 著)开源书稿:从硬件约束和模型架构出发,量化推导 LLM 推理与训练系统设计。含全书正文、PDF、配套计算工具与实验

An Open Book That Derives LLM Infra Design From Hardware Numbers — This free, open-source book works from hardware constraints and model architecture to quantitatively derive LLM training and inference system design, covering topics like KV-cache, MoE, and datacenter networking. It's a companion to the author's earlier AI Agent book and targets engineers who want to reason about latency and cost tradeoffs, not just use infra as a black box. Includes full text, PDF/EPUB, calculation tools, and experiments.

LLMs & MLacceleratorai-infraai-infrastructurebookdatacenter-networkdeepseekdistributed-systemsgpukv-cachellmllm-inferencellm-servingllm-trainingmixture-of-expertsperformance-engineering

Metrics

Relative growth (7d)+7%Growth over the last 7 days relative to size
Acceleration+118%Change in growth pace from the week before
Attention65The three metrics scaled to 0–100 and combined

Retention can be computed 30 days after a repository first trends (it isn't shown until then). See About for how each metric is defined.

Star growth and Trending appearances (last 30 days)

First listed+173★/dayDaily
Oct 22, 2026Nov 20, 2026

Bands mark days on the trending list (darker = higher rank)

Daily trending ranks as a table (1 days)
DateDaily
Nov 16, 2026
python #4 +127

Created Aug 23, 2026 · First on trending Nov 16, 2026

Other notable repositories in the same category (LLMs & ML)

See all LLMs & ML repositories →

Rank within deepseek

#Repository★
1ollama/ollama187,682
2huggingface/transformers169,298
3langgenius/dify161,533
4Snailclimb/JavaGuide159,624
5nexu-io/open-design108,555
6vllm-project/vllm97,238
7cheaper-inference/OmniRoute88,975
8lobehub/lobehub85,881
9Ebony-Vinyl/dsh-our-free-model81,200
10unslothai/unsloth79,804
25bojieli/ai-infra-book8,137

See all deepseek repositories →

Get the weekly digest by email

Every Monday: the repositories that grew most last week, and how the buzz spread between the global and Japanese-language web. Unsubscribe any time. See the privacy and ad policy for how we handle your address.