Commits

History for

Commits on Aug 25, 2023

Correct WandB log step
photomz
authored

Commits on Aug 19, 2023

new model export: versions 0 (legacy) and 1
karpathy
committed

Commits on Aug 13, 2023

bigchange: add multiquery support in run.c. we can now train and inference multiquery models (where n_kv_heads < n_heads). this also means that we, in principle, support Llama 2 34B and 70B models,…
karpathy
committed
make default be the llama2 tokenizer
karpathy
committed
remove the tinyshakespeare dataset until i can bring it back later in a nicer form, otherwise right now we just have a ton of copy paste code here
karpathy
committed

Commits on Aug 11, 2023

ok i can train and sample a model with a custom tokenizer
karpathy
committed

Commits on Aug 6, 2023

fix bug, have to use raw_model not model to access the loss
karpathy
committed

Commits on Aug 4, 2023

Updating training code for loss result
mpcusack-color
committed

Commits on Aug 1, 2023

Add tinyshakespeare dataset
wlamond
committed

Commits on Jul 27, 2023

Commits on Jul 24, 2023

Have DDP ignore `freqs_cis` to avoid broadcast
Andrew Gu
committed

Commits on Jul 23, 2023

Loading