Search results

README typo: "You'll need a cloud GPU box if you don't have enough, for that I recommend Lambda."

Status: Open.

#108 In karpathy/build-nanogpt;

· tv42 opened on Mar 11, 2026

Dataloader Out of bounds check

Status: Open.

#105 In karpathy/build-nanogpt;

· krishhrana opened on Sep 21, 2025

Error If I try to load "gpt2" with model = GPT.from_pretrained('gpt2')

Status: Open.

#103 In karpathy/build-nanogpt;

· Terrambc opened on Jul 22, 2025

Model convergence differences between cpu and mps

Status: Open.

#92 In karpathy/build-nanogpt;

· xsatvik opened on Mar 9, 2025

Saving raw_model.state_dict() checkpoints

Status: Open.

#90 In karpathy/build-nanogpt;

· anw-g01 opened on Jan 28, 2025

Why does weights sharing have to be in that direction?

Status: Open.

#89 In karpathy/build-nanogpt;

· BetterWang opened on Jan 16, 2025

How Can I extract Last Layer Representation?

Status: Open.

#86 In karpathy/build-nanogpt;

· shantanu778 opened on Oct 31, 2024

Avoid tiktoken.decode panic on unknown tokens.

Status: Open.

#81 In karpathy/build-nanogpt;

· IggShaman opened on Aug 26, 2024

torch.compile-d models do not work with example generation and hellaswag eval

Status: Open.

#79 In karpathy/build-nanogpt;

· IggShaman opened on Aug 26, 2024

Text generation can use raw_model instead of model

Status: Open.

#56 In karpathy/build-nanogpt;

· sapphire008 opened on Jul 10, 2024

Different inference results between flash attention and manually implemented attention appeared.

Status: Open.

#50 In karpathy/build-nanogpt;

· Jaeckel-d opened on Jul 2, 2024

How to support padding in the train dataset for training ?

Status: Open.

#49 In karpathy/build-nanogpt;

· mrhimanshu opened on Jul 2, 2024