D2L-TVM v0.1 is released

Dear community,

We are happy to release the first version of the TVM introduction book.

This release contains 22 sections, including:

  1. Getting Started
    1. Installation
    2. Vector Addition
    3. Neural Network Inference
    4. Running on a Remote Machine
  2. Expressions for Operators
    1. Data Types
    2. Shapes
    3. Index and Shape Expressions
    4. Reduction Operations
    5. Conditional Expression: if-then-else
    6. Truth Value Testing: all and any
  3. Common Operators
    1. Matrix Multiplication
    2. Convolution
  4. Operator Optimizations on CPUs
    1. CPU Architecture
    2. Function Call Overhead
    3. Vector Addition
    4. Matrix Multiplication
    5. Improve Cache Efficiency by Blocking
    6. Convolution
    7. Packed Convolution
  5. Operator Optimizations on GPUs
    1. GPU Architecture
    2. Vector Addition
    3. Matrix Multiplication

Per suggestions from D2L-TVM: A TVM introduction book,

  • we added a roadmap doc
  • each section has a Colab button so that it can be run on colab directly
  • we also embedded discussion threads at the end of some sections. You probably will see a lot empty threads will be created later.

Lastly, we are seeking contributors to add new chapters, please contact me if you are interested.

17 Likes