mlnlp-world/deeplearning-muli-notes

★ 3,777Jupyter NotebookAudience · researcherComplexity · 3/5Setup · moderate

Mindmap

mindmap
  root((D2L study notes))
    What it is
      73-lecture course notes
      Chinese community project
      Free learning resource
    Content
      Markdown lecture notes
      Jupyter notebooks
      PyTorch code
    Topics covered
      Linear networks and MLPs
      CNNs and ResNet
      RNNs and LSTMs
      Attention and BERT
    How to use
      Watch video on Bilibili
      Read matching notes
      Run PyTorch notebooks
    Audience
      Deep learning students
      Chinese-speaking learners

mindmap root((D2L study notes)) What it is 73-lecture course notes Chinese community project Free learning resource Content Markdown lecture notes Jupyter notebooks PyTorch code Topics covered Linear networks and MLPs CNNs and ResNet RNNs and LSTMs Attention and BERT How to use Watch video on Bilibili Read matching notes Run PyTorch notebooks Audience Deep learning students Chinese-speaking learners

Click or tap to explore — scroll the page freely

Things people build with this

USE CASE 1

Study deep learning from scratch by following 73 free video lectures with matching notes and runnable code notebooks.

USE CASE 2

Practice implementing neural network architectures including CNNs, RNNs, LSTMs, attention, and BERT using PyTorch.

USE CASE 3

Use the notes as a structured companion alongside the free Dive into Deep Learning textbook (Chinese or English).

USE CASE 4

Run the PyTorch code examples for each chapter to see how models are built, trained, and evaluated with real code.

Tech stack

PythonPyTorchJupyter Notebook

Getting it running

Difficulty · moderate Time to first run · 30min

Requires Python and PyTorch installed, some chapters need a GPU for reasonable training times.

In plain English

This repository is a community-compiled set of study notes and code for a deep learning video course taught by Mu Li, a senior scientist at AWS and a Carnegie Mellon University computer science PhD. The course, called "Dive into Deep Learning" (or in Chinese, written as the title shown in the README), consists of 73 video lectures, each kept under 30 minutes, and is freely available on the Chinese video platform Bilibili. The creators of this repository took detailed notes while working through the course and published them here so other students could benefit. What you get in this repository is two things for each lecture: a markdown note file that tracks alongside the video, and a Jupyter notebook with complete Python code and Chinese-language comments that you can run yourself. The intent is that you can watch a lecture and read the matching notes at the same time to reinforce understanding, then run the notebook to practice the concepts with actual code. The course covers deep learning from the ground up. It includes foundational topics like linear neural networks and multi-layer perceptrons, then works through more advanced architectures including convolutional networks (both classic designs like LeNet and modern ones like ResNet), recurrent networks and LSTMs, the attention mechanism, and language models including BERT. It also touches on computer vision, natural language processing, optimization methods, and recommender systems. All code in the notebooks uses PyTorch. The notes and code are organized into 19 chapters matching the textbook "Dive into Deep Learning," which exists in both a Chinese version and an English version and is linked from the README. According to the project description, working through the material during a winter break of roughly 40 days is the expected pace. All content is free. The textbook, the video lectures, and the notes and code in this repository are all openly accessible. The full README is longer than what was shown.

Copy-paste prompts

Prompt 1

I'm working through the Dive into Deep Learning course. Show me how to implement a simple multi-layer perceptron in PyTorch from scratch.

Prompt 2

Help me understand the attention mechanism chapter from the D2L course. Write a PyTorch implementation of scaled dot-product attention with comments explaining each step.

Prompt 3

I want to run the ResNet chapter code from D2L. Show me how to build a basic ResNet block in PyTorch and train it on CIFAR-10.

Prompt 4

Explain how LSTM networks work for sequence modeling and show me a PyTorch implementation similar to the D2L recurrent networks chapter.

Prompt 5

I'm following the D2L NLP chapter. Show me how to fine-tune a BERT model for text classification using PyTorch and Hugging Face.

Open on GitHub → Explain another repo

← mlnlp-world on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.