Commit History

  1. star port sidecar

    Jesse Daniel Brown
    89f6f6d
  2. star port hot-path row

    Jesse Daniel Brown
    b0c3f1b
  3. star port sidecar

    Jesse Daniel Brown
    d3d295a
  4. star port: serve raw

    Jesse Daniel Brown
    b46c93c
  5. star port sidecar

    Jesse Daniel Brown
    34630ea
  6. star port: light this stone so it is reachable

    Jesse Daniel Brown
    236390c

  1. new examples folder

    Kye Gomez
    227dbb1
  2. tiny tests

    Kye Gomez
    963e112
  3. flash attn

    Kye Gomez
    7d78ebe

  1. fix training

    Kye Gomez
    eae0f04
  2. [bugf][act-halting][gate halted positions from weight accumulation][bugf][moe-router-bias][stop⁠

    Kye Gomez
    289981b
  3. [improvement][loguru-logging][replace print with loguru in training script][feat][ckpt-logging][add checkpoint start and success log events][docs][readme-optimizer][remove muon optimizer⁠

    Kye Gomez
    7ba6907
  4. [fix][rope Every decode token was stuck at position 0, so <q_decoded, k_cached> lost the (n - m) term entirely]

    Kye Gomez
    18cca89
  5. just use adam for now in training maybe add muon later

    Kye Gomez
    537b116
  6. readme

    Kye Gomez
    137cd88
  7. readme

    Kye Gomez
    f37d405
  8. readme

    Kye Gomez
    ef2cac6
  9. [docs][readme-badges][add pypi twitter github badges to⁠

    Kye Gomez
    12f6c5b
  10. [feat][training-script][add 3b fineweb-edu training⁠

    Kye Gomez
    5ffb897
  11. use gpt-oss tokenizer because it's a great tokenizer

    Kye Gomez
    97bc414
  12. [feat][dropout][add dropout to config attn and⁠

    Kye Gomez
    5cfef74
  13. Update README.md

    Kye Gomez
    Unverified signature
    0623ceb
  14. [Update variants with parameter count equation]

    Kye Gomez
    a825ba2
  15. more references fix

    Kye Gomez
    f00d10b

  1. clean up

    Kye Gomez
    e522ecf
  2. format

    Kye Gomez
    1c54259
  3. variants

    Kye Gomez
    9cce1c1
  4. pip and readme

    Kye Gomez
    299a0cd
  5. [bugf][lora-b-init][fix zero-init B making adapter always output zero][bugf][lti-get-a][fix 0 times⁠

    Kye Gomez
    806a8da
  6. disclaimer and rope tests

    Kye Gomez
    53f786a
  7. installation

    Kye Gomez
    14a806c
  8. cleanup

    Kye Gomez
    7abd3e5
  9. readme

    Kye Gomez
    0818bf8
  10. [DOCS] [LICENSE]

    Kye Gomez
    fb2939b

  1. tests

    Kye Gomez
    0699c00
  2. example .py

    Kye Gomez
    c258cdc
  3. [FEAT][Main.py]

    Kye Gomez
    7916266
  4. moe

    Kye Gomez
    4ce503f
  5. paper

    Kye Gomez
    79b916d
  6. reference

    Kye Gomez
    4aea9d9
  7. Initial commit

    Kye Gomez
    Unverified signature
    8fa7022