Skip to content

KEP-2401: Kubeflow LLM Trainer V2 #2410

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
google-oss-prow merged 70 commits into from
Mar 11, 2025
Merged

KEP-2401: Kubeflow LLM Trainer V2 #2410

Show file tree
Hide file tree
Changes from 69 commits
Commits
Show all changes
70 commits
Select commit Hold shift + click to select a range
dc8d6bb
doc: add initial doc for KEP-2401.
Electronic-Waste Jan 31, 2025
d564618
doc: update motivation.
Electronic-Waste Feb 1, 2025
02dd56d
doc: add llm lifecycle picture.
Electronic-Waste Feb 1, 2025
a695ec5
doc: add goals and non-goals.
Electronic-Waste Feb 1, 2025
65c1d17
doc: add alternatives.
Electronic-Waste Feb 1, 2025
779024a
doc: add proposal chapter.
Electronic-Waste Feb 1, 2025
4cd104d
doc: add multiple frameworks support section in design chapter.
Electronic-Waste Feb 1, 2025
f673865
doc: add data preprocess section in design chapter.
Electronic-Waste Feb 1, 2025
e6db3ce
doc: add fine-tuning config section in design details chapter.
Electronic-Waste Feb 1, 2025
24dd0e4
fix(doc): remote all trailing whitespaces.
Electronic-Waste Feb 1, 2025
1e46874
doc: update llm-trainer-v2-workflow img.
Electronic-Waste Feb 18, 2025
86e10b3
doc: update goals and non-goals.
Electronic-Waste Feb 18, 2025
75e8494
doc: remove torchrun proposal to alternatives.
Electronic-Waste Feb 18, 2025
1b698ec
doc: move torchrun design to alternatives.
Electronic-Waste Feb 18, 2025
618db86
doc: move some fine-tuning config not support by torchtune to alterna…
Electronic-Waste Feb 18, 2025
ed502c6
doc: move torchtune sections to proposal and design chapters.
Electronic-Waste Feb 18, 2025
49f5f02
doc: update proposal & move FSDP config to alternatives.
Electronic-Waste Feb 19, 2025
8e5f256
doc: update fine-tuning config & unify lora/qlora/dora.
Electronic-Waste Feb 19, 2025
bdc4ef7
fix(doc): update fine-tuning config & fix doc according to comments.
Electronic-Waste Feb 19, 2025
30c8b84
doc: add model & dataset initialization / model exporting.
Electronic-Waste Feb 19, 2025
11f9cc4
doc: add dataset preprocess/tokenizer chapter.
Electronic-Waste Feb 19, 2025
7a0411c
doc: fix some errors in doc.
Electronic-Waste Feb 19, 2025
86033ae
doc: update chapter name.
Electronic-Waste Feb 19, 2025
3a39c2b
doc: add type in the diagrams.
Electronic-Waste Feb 19, 2025
ae15088
doc: add optimizer and scheduler config.
Electronic-Waste Feb 19, 2025
a19603e
fix(doc): fix some errors in doc.
Electronic-Waste Feb 19, 2025
935bb67
doc: add initial parameter override.
Electronic-Waste Feb 19, 2025
e910167
doc: update config override.
Electronic-Waste Feb 19, 2025
39a54da
doc: fix some errors in doc.
Electronic-Waste Feb 19, 2025
8d29318
fix(doc): add CustomTrainingConfig dataclass.
Electronic-Waste Feb 20, 2025
67bd3b7
fix(doc): integrate torchtune mutation logic into torch plugin.
Electronic-Waste Feb 21, 2025
e946c19
fix(doc): split torchtune config chapter.
Electronic-Waste Feb 23, 2025
2b1bccc
fix(doc): add two options for SDK & seperate LoRA chapter.
Electronic-Waste Feb 23, 2025
f4fee07
doc: add an example to show parameters mutation.
Electronic-Waste Feb 25, 2025
b16de4d
doc: add detailed design on mutation in torch plugin.
Electronic-Waste Feb 25, 2025
7605cb8
doc: add dir structure for option 1.
Electronic-Waste Feb 25, 2025
b84a61b
doc: add dir structure for option 2.
Electronic-Waste Feb 25, 2025
a2eb388
doc: add Test Plans chapter.
Electronic-Waste Feb 25, 2025
a9ac42e
fix(doc): remove device parameter.
Electronic-Waste Feb 25, 2025
9d95b75
fix(doc): fix typo error.
Electronic-Waste Feb 25, 2025
54ab01e
fix(doc): fix code line format.
Electronic-Waste Feb 25, 2025
99c8a56
fix(doc): update error in proposal example & add num_nodes and resour…
Electronic-Waste Feb 25, 2025
35438f3
fix(doc): update manifests dir in option 1.
Electronic-Waste Feb 26, 2025
95e0465
fix(doc): split complement torch plugin chapter.
Electronic-Waste Feb 26, 2025
82ee2c0
fix(doc): move option 1 (reserving recipe and config) to alternatives…
Electronic-Waste Feb 27, 2025
fc239cd
fix(doc): update goals & add description in propagate torchtune setti…
Electronic-Waste Feb 28, 2025
ce66249
chore(doc): complete map section in SDK.
Electronic-Waste Feb 28, 2025
b62c2d5
chore(doc): add maintaining ClusterTrainingRuntime chapter.
Electronic-Waste Feb 28, 2025
cbe1800
fix(doc): update recipe selection.
Electronic-Waste Feb 28, 2025
e9e2ddb
fix(doc): remove some CTRs & only reserve llama family.
Electronic-Waste Mar 4, 2025
ba831d3
fix(doc): rename TorchtuneConfig to TorchTuneConfig.
Electronic-Waste Mar 4, 2025
4df02b3
fix(doc): remove name prefix in CTRs.
Electronic-Waste Mar 4, 2025
ec057ea
fix(doc): update TrainJob and CTR example.
Electronic-Waste Mar 4, 2025
4e5d414
fix(doc): fix some typos & address comments.
Electronic-Waste Mar 4, 2025
9bed32c
fix(doc): update complement torch plugin section.
Electronic-Waste Mar 4, 2025
4842b07
fix(doc): add gemma2 mistral qwen2_5 back.
Electronic-Waste Mar 5, 2025
e0e474c
fix(doc): update implementation history.
Electronic-Waste Mar 5, 2025
601ef3a
fix(doc): remove the name prefix in CTRs.
Electronic-Waste Mar 5, 2025
c3c1bb8
fix(doc): update typo according to the review.
Electronic-Waste Mar 5, 2025
abcb00d
chore(doc): add webhook section.
Electronic-Waste Mar 6, 2025
7133ae4
chore(doc): add webhook func description.
Electronic-Waste Mar 6, 2025
280831b
fix(doc): update item format.
Electronic-Waste Mar 6, 2025
eadbe73
fix(doc): add the lifecyle of LLM fine-tuning with torchtune.
Electronic-Waste Mar 6, 2025
b48cd72
fix(doc): remove diagram description.
Electronic-Waste Mar 6, 2025
781d4d3
fix(doc): reorg and update the doc according to the review.
Electronic-Waste Mar 6, 2025
14332e7
fix(doc): fix some typos.
Electronic-Waste Mar 6, 2025
1726a5e
fix(doc): fix some format error.
Electronic-Waste Mar 6, 2025
b3d2b6f
fix(doc): update implementation history.
Electronic-Waste Mar 6, 2025
bfc48fa
fix(doc): rename CTRs' file name.
Electronic-Waste Mar 7, 2025
29808b7
fix(doc): remove detailed design.
Electronic-Waste Mar 11, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Loading