Skip to content

Pull requests: karpathy/autoresearch

New pull request New
145 Open 253 Closed
145 Open 253 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

feat(train): GPU-aware MFU, warmup MFU fix, experiment integrity log
#623 opened Jun 27, 2026 by eli-labz Loading…
feat: support running agent and training process on same gpu
#621 opened Jun 23, 2026 by mohamed-em2m Loading…
feat(novelty): add experiment novelty score checker to prevent redundant runs
#619 opened Jun 21, 2026 by aniruddhaadak80 Loading…
1
1
feat(train): add optional PyTorch Profiler wrapper outputting LLM-readable summary
#618 opened Jun 21, 2026 by aniruddhaadak80 Loading…
1
feat(cloud): add setup-cloud.sh and document root-only cloud setups in README
#617 opened Jun 21, 2026 by aniruddhaadak80 Loading…
1
fix(prepare): harden tokenizer and token_bytes cache loading with sha256 checksums
#616 opened Jun 21, 2026 by aniruddhaadak80 Loading…
1
fix(train): dynamically detect GPU capability and look up peak FLOPS for MFU calculation
#615 opened Jun 21, 2026 by aniruddhaadak80 Loading…
1
1
feat: add CLI analysis tool for experiment results and agent feedback
#611 opened Jun 13, 2026 by aniruddhaadak80 Loading…
1
5
fix(prepare): make Tokenizer.encode prepend encoding robust for lists, tuples, and strings
#610 opened Jun 13, 2026 by aniruddhaadak80 Loading…
1
3
fix(train): save model checkpoint before evaluation to prevent losing results on eval crash
#609 opened Jun 13, 2026 by aniruddhaadak80 Loading…
1
4
fix(prepare): verify cache size integrity against server Content-Length
#608 opened Jun 13, 2026 by aniruddhaadak80 Loading…
1
5
fix(prepare): use true raw bytes for BPB token byte count calculation
#607 opened Jun 13, 2026 by aniruddhaadak80 Loading…
1
3
fix(pyproject): specify py-modules to fix setuptools editable install discovery
#606 opened Jun 13, 2026 by aniruddhaadak80 Loading…
1
1
fix(tokenizer): support multi-token prepend and add LICENSE (#553, #575)
#605 opened Jun 13, 2026 by aniruddhaadak80 Loading…
10
fix(train,prepare): resolve open issues (#556, #547, #549, #552, #542)
#604 opened Jun 13, 2026 by aniruddhaadak80 Loading…
8
Add research orchestrator and enhance cross-platform support
#597 opened May 30, 2026 by exdsgift Loading…
3 tasks done
Add non-CUDA runtime fallback and documentation
#595 opened May 22, 2026 by arunshar Loading…
Bump PyTorch to 2.11.0 (fixes silent expandable_segments bug in 2.9.1)
#588 opened May 11, 2026 by srosro Loading…
4 tasks done
2
feat(mps): Apple Silicon MPS port — Flash Attention 3 → PyTorch SDPA
#582 opened May 10, 2026 by Rensp Loading…
3 tasks
feat: add Apple Intel CPU + MPS support
#576 opened May 4, 2026 by amul3 Loading…
Add MIT license file matching README
#575 opened May 3, 2026 by hummbl-dev Loading…
fix(train): size rotary cache to config.sequence_len, not 10x
#557 opened Apr 30, 2026 by lonexreb Loading…
2 of 3 tasks
1
fix(train): warmup off-by-one in steady_state_mfu accumulation
#556 opened Apr 30, 2026 by lonexreb Loading…
3 tasks done
fix(tokenizer): support multi-token prepend and tighten scope (#348)
#553 opened Apr 30, 2026 by lonexreb Loading…
8 tasks done
1
fix(prepare): use raw bytes for BPB to avoid U+FFFD inflation (#384)
#552 opened Apr 30, 2026 by lonexreb Loading…
1 of 3 tasks
1
ProTip! What’s not been updated in a month: updated:<2026-05-28.