Issues

• test

  karpathy/nanoGPT#751 · Filter by author hermes-agent-jx opened on Aug 10, 2026

• Realtime response generation simulation instead of one big response at the end.

  karpathy/nanoGPT#748 · Filter by author prethiv opened on Jul 31, 2026

• GPT.generate() crashes on temperature=0 instead of doing greedy decoding

  karpathy/nanoGPT#738 · Filter by author NikhilSKashyap opened on Jun 17, 2026
  1

• Code quality scan: 9 findings (C-, 68/100)

  karpathy/nanoGPT#731 · Filter by author repobilitycom opened on May 16, 2026

• the model is not domain agnostic

  karpathy/nanoGPT#722 · Filter by author br3eze-code opened on Apr 19, 2026

• Critical Sandbox Escape Vulnerability in nanochat

  karpathy/nanoGPT#717 · Filter by author captain99hook opened on Apr 16, 2026

• Qodo AntiSlop scan found 8 issues across 7 recent PRs

  karpathy/nanoGPT#708 · Filter by author gvago opened on Apr 4, 2026

• small warning

  karpathy/nanoGPT#704 · Filter by author Damian-o2 opened on Mar 31, 2026
  1

• Scalar C++ Transformer Implementation — Verified Against PyTorch

  karpathy/nanoGPT#693 · Filter by author dratman opened on Mar 12, 2026

• Regarding your LLMOS concept

  karpathy/nanoGPT#692 · Filter by author darthcoder opened on Mar 11, 2026

• Why not just surpass 5.2 directly? Let's add the LiE enhancement plugin.

  karpathy/nanoGPT#686 · Filter by author ZengHaohei opened on Feb 18, 2026

• Meet TrueNanoGPT

  karpathy/nanoGPT#685 · Filter by author nanocreator2 opened on Feb 15, 2026

• Downstream works

  karpathy/nanoGPT#677 · Filter by author Mec-iS opened on Jan 6, 2026

• CustomV2: A New Activation Function That Wins Under Early Stopping

  karpathy/nanoGPT#675 · Filter by author vbepipe opened on Jan 3, 2026

• Feature request: Would it be possible to make a nano example of MoE models?

  karpathy/nanoGPT#665 · Filter by author jzju opened on Nov 18, 2025

• [Security] Code Execution via unsafe deserialization

  karpathy/nanoGPT#661 · Filter by author edoardottt opened on Nov 13, 2025

• Dataloader segmentation fault

  karpathy/nanoGPT#656 · Filter by author mtodescato opened on Oct 31, 2025

• MPS + float16: GradScaler / autocast behavior in train.py

  karpathy/nanoGPT#654 · Filter by author v659 opened on Oct 23, 2025

• Feature request: semantic audit during training loop

  karpathy/nanoGPT#653 · Filter by author elly99-AI opened on Oct 22, 2025

• Enable Private Security Reporting

  karpathy/nanoGPT#646 · Filter by author edoardottt opened on Oct 16, 2025

• Use Nsight (nsys profile) to profile the training in a single multi-GPU node

  karpathy/nanoGPT#642 · Filter by author Minyuluck opened on Oct 1, 2025

• GUFF

  karpathy/nanoGPT#636 · Filter by author si-open opened on Sep 3, 2025

• FutureWarning: torch.cuda.amp.GradScaler(args...) is deprecated.

  karpathy/nanoGPT#635 · Filter by author si-open opened on Sep 3, 2025

• RuntimeError: Dataset scripts are no longer supported

  karpathy/nanoGPT#633 · Filter by author mlang opened on Aug 19, 2025
  5

• Sharing Learn-nanoGPT: An Educational NanoGPT Version with Hardware-Adaptive Training for Shakespeare Text

  karpathy/nanoGPT#631 · Filter by author niloydebbarma-code opened on Aug 16, 2025