01:33:07

Using The C++ `unordered_map` In Spiral

01:48:02

Adding The `stack_refs` And `heap_refs` Layout Types To The Spiral Language

04:39:59

Drawing From A Deck In Leduc On The Host

32:50

How To Build Cuda Files In VS Code

01:23:21

Difficulties Translating The Leduc Game From C++ To Python

03:59:23

Sketching Out The HD-CFR Algorithm Using Numpy

05:45

Intro To The Hopfield Dictionary CFR Playlist

49:33

Associative Memory CFR Could Be A Viable Method (rant, no programming)

01:22:31

Cross Entropy Method Training On HU No Limit Holdem

04:59:16

Cross Entropy Method Training On Leduc Poker

02:10:37

Redesigning Spiral's ML Library To Support Arbitrary Parameter Spaces

01:19:22

Sketching Out The Cross Entropy Method In Code

01:17:00

Implementing A Transpose Kernel On The GPU In The Spiral Language

02:56:31

Counterfactual Regret Training For Heads Up No Limit Hold'em (Using A Random NN Model)

03:27:48

Adding Stack Mutable Types To The Spiral Language

01:43:02

Revisiting Ampere Matrix Multiplication

01:59:09

Revisiting Ampere Async Loads

01:31:10

Optimizing The Training Loop For The GPU Implementation Of The CFR Algorithm

01:31:51

Training A Superhuman Leduc Agent On The GPU Using The CFR Algorithm

01:01:03

Replacing The Use Of Static Shared Memory With Dynamic In Spiral's ML Library

06:06:40

Parallel Tabular CFR Training On The GPU

41:08

Improving The ML Library Testing. We're Going Back To Timelapsing.

06:46:32

Integrating Tabular CFR With The Poker Games On The GPU

02:32:01

Debugging The Parallel Tabular CFR Implementation (Part 2)

03:44:33

Sharing The Pointers In Shared Memory Automatically Using The Transposing Loops

02:01:22

Debugging The Parallel Tabular CFR Implementation (Part 1)

08:54:41

Sketching Out A Parallel Tabular CFR Algorithm For The GPU

06:51:57

Creating The `row_gather` Primitives

05:45

How To Do Mouse Highlighting On Windows 11

07:19:23

Running The Game Model With All The Threads. Implementing The Neural Model Loop.

04:31:35

Running The (Random) Neural Agent On NL Holdem

03:56:30

Running The (Random) Neural Agent On Leduc

06:01:58

Refactoring The ML Library (Part 3)

06:52:07

Refactoring The ML Library (Part 2)

04:19:34

Adding Higher Ranked Types To Spiral

04:09:08

Implementing The Masked Softmax

11:21:56

Refactoring The ML Library (Part 1)

06:10:20

Adding ECharts To The Game Frontend

17:17

Fixing A GADT Error In The Spiral Language

07:41:55

Redesigning Spiral's Machine Learning Library With GADTs

01:32:24

Review Of GADTs In Spiral And Their Implementation (livestream)

04:37:24

Implementing GADTs (Generalized Algebraic Datatypes) In Spiral (Timelapse)

01:16:59

Pickling The Leduc Poker Game's State Into The ML Model

02:58:49

Integrating The ML Cuda Primitives Into Spiral's ML Library

15:59

Getting Started With Cutlass On Windows And Visual Studio

04:18:25

Creating The Fully Fledged Cuda Backend With C++ Style Reference Counting For Spiral

02:18:01

Failed Attempt At Making A Cuda Backend With C Style Reference Counting...

03:10:25

Making A Fully Fused ML Library In Spiral (Part 1)

03:43:25

Implementing The Heads Up No Limit Hold'em Game On The GPU

02:50:03

Implementing The Leduc Poker Game On The GPU

02:46:32

Implementing The Leduc Poker Game On The Host

59:09

Implementing The RPS Game On The GPU (Part 2)

47:10

Serializing JSON Formatted Python Objects In Spiral

01:25:41

Serialization Of Static And Resizable Arrays in Spiral

05:21:53

Extending Spiral's Tokenizer To Allow Parsing Of Macro Expressions In Parens (feat. Fparsec)(stream)

01:31:47

Implementing the RPS game on the GPU (Part 1)

01:49:16

Doing The Frontend And The Backend For The RPS Game In Python

58:30

Setting up a Typescript + Lit + Flask + SocketIO + ESBuilt + Poetry project

02:34:36

Pickler Combinators And Serialization In Spiral

01:35:33

Ghostlike's Cross Entropy Method (long rant, no programming)