Morph MORPH
SIGN IN SIGN UP
karpathy / nanoGPT UNCLAIMED

The simplest, fastest repository for training/finetuning medium-sized GPTs.

0 0 0 Python

COMMITS

master
November 12, 2025
A
Update README to mention nanochat and deprecation
Andrej committed
3adf61e
December 9, 2024
A
Merge pull request #578 from devin-open-source/devin/1733728337-fix-warmup-lr
Andrej committed
93a43d9
D
fix: ensure non-zero learning rate during warmup at iteration 0
Devin AI committed
dc81cb3
June 3, 2024
A
Merge pull request #463 from goswamig/test1
Andrej committed
9755682
A
Merge branch 'master' into test1
Andrej committed
3ab86ce
A
Merge pull request #487 from jellehak/patch-1
Andrej committed
7c7e627
May 28, 2024
J
Update README.md
Jelle Hak committed
5cb16fe
March 24, 2024
G
Fixing eval path in README
Gautam Kumar committed
1ab9ec1
February 27, 2024
A
Merge pull request #420 from vinjn/fix-371-enc-is-not-defined
Andrej committed
325be85
A
Merge pull request #429 from adambala/fixes
Andrej committed
a022d02
A
Merge pull request #428 from kjslag/memmap-memory-leak
Andrej committed
f68ac22
January 25, 2024
A
fix: prepare.py - added input file opening in UTF-8 encoding
Adam Isakov committed
f35dc82
A
feature: .gitignore - added venv folders
Adam Isakov committed
b7e194a
K
fix np.memmap memory leak
Kevin Slagle committed
5156fef
January 12, 2024
V
Move enc to gloabal namespace
vinjn committed
dccf362
June 22, 2023
A
Merge pull request #309 from ho2103/master
Andrej committed
eba36e8
June 19, 2023
O
Fix AssertionError on macOS - need to check CUDA availability for bf16
o committed
1eaceae
June 18, 2023
A
Merge pull request #305 from okuvshynov/fix_osx_dataload
Andrej committed
4eb7a96
O
nanogpt: fix multiprocessing in load_dataset on os x
Oleksandr Kuvshynov committed
542ac51
June 17, 2023
A
Merge pull request #301 from okuvshynov/master
Andrej committed
41d7014
O
nanogpt: allow multithreading in load dataset
Oleksandr Kuvshynov committed
bb7e967
June 14, 2023
A
use WORLD_SIZE instead of device_count, supports both the case where the number of gpus we train on is smaller than gpus available, and also multinode training may be a bugfix
Andrej Karpathy committed
7339b90
A
Merge pull request #274 from apivovarov/gelu
Andrej committed
f08abb4
A
Merge pull request #275 from apivovarov/rm_unsqueeze
Andrej committed
18ee6b6
A
Merge pull request #270 from LaihoE/master
Andrej committed
ed7887c
A
Merge pull request #276 from apivovarov/gitign
Andrej committed
8020bb5
A
Merge pull request #277 from apivovarov/is_bf16_supported
Andrej committed
0f06d9b
A
Merge pull request #286 from ctjlewis/master
Andrej committed
cf4835e
June 1, 2023
L
docs: simplify dependencies installation
Lewis committed
eeac873
May 17, 2023
A
Use bf16 only if supported
Alexander Pivovarov committed
eb33b8b