Code

History

Commits on move/encoder_kernel70+
  1. Close fileAleksa Gordic committed 2 years agodf0a826
  2. Fix mem leakAleksa Gordic committed 2 years ago30d81cc
  3. Use local params for num blocksAleksa Gordic committed 2 years ago79cbc03
  4. Add clarification on Box-MullerAleksa Gordic committed 2 years agoae9c957
  5. fix compilation with older nvccErik Schultheis committed 2 years agoa902323
  6. Remove redundant CPU computationAleksa Gordic committed 2 years ago290c00a
  7. Merge pull request #495 from ChrisDryden/shared_memoryAndrej committed 2 years agoa26041a
  8. Merge pull request #509 from ngc92/inplaceAndrej committed 2 years ago9cf8c2f
  9. Merge pull request #511 from gordicaleksa/gordicaleksa_fix_docsAndrej committed 2 years agobd0f036
  10. Update compile cmd in the dev/cuda READMEAleksa Gordic committed 2 years ago8eafd40
  11. rename gelu_backward to gelu_backward_inplace, and made it obvious in the code that this is an inplace operationErik Schultheis committed 2 years ago761fb2c
  12. added missing checksErik Schultheis committed 2 years agocc4bed0
  13. fix recompute bugAndrej Karpathy committed 2 years ago9c8edaf
  14. fix bug in resumption of dataloader, have to reload shardAndrej Karpathy committed 2 years ago0bc24c3
  15. fix argparse comment tooAndrej Karpathy committed 2 years ago1de3bef
  16. Merge branch 'ChrisDryden-layernorm_backwards_mem_saving'Andrej Karpathy committed 2 years ago60a1037
  17. make things a bit more consistent, keep recompute 1 as default, and add a commentAndrej Karpathy committed 2 years agof4874a9
  18. Merge branch 'layernorm_backwards_mem_saving' of https://github.com/ChrisDryden/llm.c into ChrisDryden-layernorm_backwards_mem_savingAndrej Karpathy committed 2 years ago629a240
  19. fix bug in readmeAndrej Karpathy committed 2 years agoed37d92
  20. refactor the README file because it was getting crazyAndrej Karpathy committed 2 years agob48224c
  21. more docs on scriptsAndrej Karpathy committed 2 years ago0fddf06
  22. also move pyrun script to scripts/Andrej Karpathy committed 2 years agofe36e40
  23. code refactor, introduce llmc dir and scripts dirAndrej committed 2 years agod98d3ed
  24. delete these files, they are now in scripts/ folder and betterAndrej Karpathy committed 2 years ago22d1607
  25. adjust readme to new scripts/ dirAndrej Karpathy committed 2 years agofc40ffc
  26. scripts folderAndrej Karpathy committed 2 years agoa0d5dbb
  27. more refactorAndrej Karpathy committed 2 years ago35aa110
  28. Merge branch 'master' into feature/libdirAndrej Karpathy committed 2 years agof8f9e49
  29. further hide the c pytorch bridge logicAndrej Karpathy committed 2 years ago5f54760
  30. bugfix don't train on val batchAndrej Karpathy committed 2 years agoca27eb8
  31. modified the recalculation parameter to be based off of 2 instead of 1Christopher committed 2 years ago59432fe
  32. Rearranged layernorm forward for 1 and 2 and used layer variables instead. Validated with test fileChristopher committed 2 years agoe5719e1
  33. First step in creating option to recalculate layernorm activations in backwards passChristopher committed 2 years ago95b1fad
  34. Merge pull request #496 from ChrisDryden/patch-7Andrej committed 2 years ago0e22982
  35. fix default number of gpus from 4 to 8 for consistency with run124M.shAndrej Karpathy committed 2 years ago862c957

Keyboard shortcuts

General

Show keyboard shortcuts
?
Focus the current search
/

Issues and pull requests

Open filters
f

Issues, pull requests and Inbox

Select next item
j↓
Select previous item
k↑
Open selected item
Enter
Dismiss filters or clear selection
Esc

Repository code

Go to file
t