- Published on
最全面的 Transformer 中文教程：从注意力直觉到 Q/K/V 手算，从多头注意力到位置编码，从 Encoder-Decoder 到 PyTorch 实现。涵盖 GPT-4/Claude/DeepSeek/LLaMA 架构对比、LLM 使用技巧的 Transformer 原理解析、2025-2026 前沿趋势（FlashAttention、Mamba/SSM、MoE），附 10 个交互式可视化和完整代码。
Pytorch
All Posts
- machine-learning (10)
- python (5)
- deep-learning (4)
- ai (4)
- tensorflow (3)
- neural-networks (3)
- llm (3)
- keras (2)
- collaborative-filtering (2)
pytorch (2)
- anomaly-detection (1)
- gaussian-model (1)
- data-science (1)
- linear-regression (1)
- gradient-descent (1)
- logistic-regression (1)
- multiclass-classification (1)
- model-evaluation (1)
- overfitting (1)
- regularization (1)
- recommender-system (1)
- decision-tree (1)
- information-gain (1)
- reinforcement-learning (1)
- deep-q-learning (1)
- dqn (1)
- experience-replay (1)
- target-network (1)
- openai-gym (1)
- lunar-lander (1)
- langchain (1)
- langgraph (1)
- ai-agent (1)
- reflection (1)
- recommender-systems (1)
- movielens-dataset (1)
- coding-agent (1)
- context-window (1)
- reasoning-vs-coding (1)
- fine-tuning (1)
- lora (1)
- qlora (1)
- adapter (1)
- huggingface (1)
- github (1)
- guide (1)
- clustering (1)
- image-processing (1)
- k-means (1)
- unsupervised-learning (1)
- writings (1)
- career (1)
- mysql (1)
- index (1)
- performance (1)
- database (1)
- openclaw (1)
- telegram-bot (1)
- devops (1)
- self-hosted (1)
- transformer (1)
- attention (1)
- self-attention (1)
- bert (1)
- gpt (1)
- nlp (1)
- java (1)
- Published on
一篇从零开始的完整教程：从理解HuggingFace推理、预训练LLM、全量微调，到参数高效微调方法（Adapter、LoRA、QLoRA）的深度解析。包含大量PyTorch代码实现、数学原理和对比实验。