🛞
设计需求

用户5190

2024年7月24日修改

画板

主课程大纲

•
词向量和大语言模型（Word2vec, GloVe）​

•
Python Review Session(Setup, Language Basics, Common Data Structures, NumPy)​

•
反向传播和神经网络基础(matrix calculus, the backpropagation algorithm)​

•
Dependency Parsing(Explicit linguistic structure and how a neural net can decide it)​

•
PyTorch Tutorial Session(Word Window Classification, Basic GANs)​

•
循环神经网络(Language Modeling, RNNs)​

•
序列到序列模型和机器翻译(Exploding and vanishing gradients, LSTMs, Bidirectional and multi-layer RNNs)​

•
LLM 简介(Machine Translation, Attention)​

•
Transformers(Impact, attention-based, Drawbacks and Variants)​

•
Pretraining and Post-training (Encoder, Decoder, RLHF, SFT, DPO)​

•
Hugging Face Transformers Tutorial Session(the Hugging Face Transformers Python library and some common patterns)​

•
基准测试与评估(Different reasons for measuring performance, Automatic Evaluation, Human Evaluation)​

•
高效的神经网络训练(DDP, FSDP, LoRA)​

•
语音脑机接口(Brain-Computer Interface)​

•
推理与代理(Reasoning, Agents)​

•
DPO 之后的生活(RLHF, RewardBench)​

•
ConvNets, 树递归神经网络和成分句法(CNNs)​

•
负责任的 NLP 简介(An Introduction to Responsible NLP)​

•
NLP、语言学和哲学(Major ideas of CS224N)​