Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Dec 3, 2024

• Update README.md

  eplarocco

  authored
  259c79e

  Copy full SHA for 259c79e

Commits on Nov 15, 2024

• added definition to notebook

  eplarocco

  committed
  e00f92d

  Copy full SHA for e00f92d

• added credit to notebook

  eplarocco

  committed
  3b10a23

  Copy full SHA for 3b10a23

• final updates _ changed imput text to trump speeches

  eplarocco

  committed
  1555f7f

  Copy full SHA for 1555f7f

Commits on Nov 13, 2024

• working model

  eplarocco

  committed
  350c194

  Copy full SHA for 350c194

• new input output

  eplarocco

  committed
  35c320f

  Copy full SHA for 35c320f

• edits

  eplarocco

  committed
  286684a

  Copy full SHA for 286684a

• edited to include onion data

  eplarocco

  committed
  318e5b3

  Copy full SHA for 318e5b3

Commits on Feb 7, 2023

• Merge pull request karpathy#9 from Andrei-Aksionov/feature/shape_description_fix

  Show description for 5220142
  karpathy

  authored
  5220142

  Copy full SHA for 5220142

• Clarify shape descriptions inside forward method

  Andrei-Aksionov

  committed
  4c8e902

  Copy full SHA for 4c8e902

Commits on Feb 6, 2023

• adding this even though the first gpt video didn't cover it, because init is very important to good performance. will cover it in followup video in more detail

  karpathy

  committed
  d38c865

  Copy full SHA for d38c865

Commits on Jan 31, 2023

• soften the wording a bit to not scare people too much

  karpathy

  committed
  5301c27

  Copy full SHA for 5301c27

• attach a note about init

  karpathy

  committed
  dfc4ece

  Copy full SHA for dfc4ece

• Merge branch 'master' of https://github.com/karpathy/ng-video-lecture

  karpathy

  committed
  ddf200a

  Copy full SHA for ddf200a

• rename the class it's a remnant of the past

  karpathy

  committed
  d5ae96f

  Copy full SHA for d5ae96f

Commits on Jan 27, 2023

• Merge pull request karpathy#6 from vineetm/master

  Show description for 17115f7
  karpathy

  authored
  17115f7

  Copy full SHA for 17115f7

Commits on Jan 25, 2023

• scale by head size

  

  Vineet Kumar

  committed
  988848a

  Copy full SHA for 988848a

Commits on Jan 17, 2023

• small readme file

  karpathy

  committed
  381c613

  Copy full SHA for 381c613

• rename v2 to gpt hah

  karpathy

  committed
  3fdc1ea

  Copy full SHA for 3fdc1ea

• might as well add these too

  karpathy

  committed
  d3df5cf

  Copy full SHA for d3df5cf

Commits on Jan 16, 2023

• parameter count

  karpathy

  committed
  e93a8c5

  Copy full SHA for e93a8c5

• add dropout, slight refactoring of variables

  karpathy

  committed
  482b15d

  Copy full SHA for 482b15d

• and add layernorms, down to 2.06

  karpathy

  committed
  0016836

  Copy full SHA for 0016836

• introduce the residual block

  karpathy

  committed
  5c3a2d2

  Copy full SHA for 5c3a2d2

• add a computation block, down to 2.24

  karpathy

  committed
  97dd3f9

  Copy full SHA for 97dd3f9

• multi-headed attention, down to 2.28 now

  karpathy

  committed
  a6e0bee

  Copy full SHA for a6e0bee

• add a single self-attention head, reach 2.5 val loss

  karpathy

  committed
  10024b1

  Copy full SHA for 10024b1

• add positional embeddings too

  karpathy

  committed
  28e5fd7

  Copy full SHA for 28e5fd7

• indirectionv2

  karpathy

  committed
  8050fde

  Copy full SHA for 8050fde

Commits on Jan 15, 2023

• v2 starter

  karpathy

  committed
  28ef287

  Copy full SHA for 28ef287

• bigram model

  karpathy

  committed
  83f7d22

  Copy full SHA for 83f7d22