tocmo0nlord/axolotl
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity
2,043 Commits 319 Branches 32 Tags
f19694131518545ac71002de01366d9b57fb894e
Commit Graph

2043 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Wing Lian
f196941315 additional fixes for docker and saving compressed 2025-04-28 13:16:29 -04:00
Rahul Tuli
5be047ac46 Fix: Test 
Signed-off-by: Rahul Tuli <rtuli@redhat.com>
 2025-04-28 13:16:29 -04:00
Rahul Tuli
758115b8c6 Apply patch from @winglian 
Signed-off-by: Rahul Tuli <rtuli@redhat.com>
 2025-04-28 13:16:29 -04:00
Rahul Tuli
0dc1da5876 Add: line about further optimizations using llmcompressor 
Signed-off-by: Rahul Tuli <rtuli@redhat.com>
 2025-04-28 13:16:29 -04:00
Rahul Tuli
f3e876dbfc Address Review Comments: 
* deleted redundant docs/llm_compressor.qmd
* incorporated feedback in integration README.md
* added llmcompressor integration to docs/custom_integrations.qmd

Signed-off-by: Rahul Tuli <rtuli@redhat.com>
 2025-04-28 13:16:29 -04:00
Rahul Tuli
99c13ef60c Add: .qmd file 2025-04-28 13:16:29 -04:00
Rahul Tuli
2c24434ee0 Tests, Style, Updates 2025-04-28 13:16:29 -04:00
Rahul Tuli
586268a0d7 Rebase and updates! 2025-04-28 13:16:29 -04:00
Rahul Tuli
b600e119b6 Add: llm_compressor integration documentation 2025-04-28 13:16:29 -04:00
Rahul Tuli
a8e5ba000e Move: LLMCompressorPlugin into it's own submodule 2025-04-28 13:16:29 -04:00
Rahul Tuli
bc3dfa666d Update model config 2025-04-28 13:16:29 -04:00
Rahul Tuli
4371f3459e Use: absolute import 2025-04-28 13:16:29 -04:00
Rahul Tuli
cc58d5e072 Rename: sft.yaml to sparse-finetuning.yaml 2025-04-28 13:16:29 -04:00
Rahul Tuli
d197b054e3 Add: llcompressor installable 2025-04-28 13:16:29 -04:00
Rahul Tuli
7e1e153831 Address review comments from @markurtz 2025-04-28 13:16:29 -04:00
Rahul Tuli
42de3096cf Apply suggestions from @markurtz 
Co-authored-by: Mark Kurtz <mark.j.kurtz@gmail.com>
 2025-04-28 13:16:29 -04:00
Rahul Tuli
27758840a1 Update llmcompressor version to latest 2025-04-28 13:16:29 -04:00
Rahul Tuli
8dbf5c215a Revert: TODO's 2025-04-28 13:16:29 -04:00
Rahul Tuli
6411ca3fe1 Use: warning over warn 2025-04-28 13:16:29 -04:00
Rahul Tuli
813809c54d pre commit hooks 2025-04-28 13:16:29 -04:00
Rahul Tuli
af7cfdc30b Add:llmcompressor instalable 2025-04-28 13:16:29 -04:00
Rahul Tuli
b76d2d1130 Update: review comments! 2025-04-28 13:16:29 -04:00
Rahul Tuli
7946f89df4 Add: SFTPlugin with llmcompressor 2025-04-28 13:16:29 -04:00
Dhruv Mullick
8b33ae1c4f Fix bug in grpo reward module import (#2571) 2025-04-28 00:31:56 -04:00
Wing Lian
dc4da4a7e2 update trl to 0.17.0 (#2560) 
* update trl to 0.17.0

* grpo + vllm no longer supported with 2.5.1 due to vllm constraints

* disable VLLM_USE_V1 for ci

* imporve handle killing off of multiprocessing vllm service

* debug why this doesn't run in CI

* increase vllm wait time

* increase timeout to 5min

* upgrade to vllm 0.8.4

* dump out the vllm log for debugging

* use debug logging

* increase vllm start timeout

* use NVL instead

* disable torch compile cache

* revert some commented checks now that grpo tests are fixed

* increase vllm timeoout back to 5min
 2025-04-27 19:19:53 -04:00
Wing Lian
f9c7c3bb72 don't use is_main_process during config validation (#2569) 2025-04-26 14:14:52 -04:00
Wing Lian
caf5cb63ea add e2e smoke test for using activation/gradient checkpointing with offload (#2565) 
* add e2e smoke test for using activation/gradient checkpointing with offload

* disable duplicate code check for the test

* fix relative import

* seq len too small to test this dataset with packing

* Fix checkpoint ptaching for tests
 2025-04-25 21:11:17 -04:00
Wing Lian
5dba5c82a8 fix support for wandb run_name for rl trainers (#2566) [skip ci] 
* fix support for wandb run_name for rl trainers

* prefer to use wandb random names for run_name
 2025-04-25 21:10:54 -04:00
Chiwan Park
e3c9d541a7 fix: crash when pretraining_dataset with dispatch_batches is false (#2558) 2025-04-25 17:15:03 -04:00
NanoCode012
9eba0ad118 chore(doc): update docker tags on doc (#2559) [skip ci] 2025-04-25 17:14:48 -04:00
Wing Lian
53dbf97d85 make cce default to true when using the plugin (#2562) [skip ci] 2025-04-25 17:14:26 -04:00
Eko Julianto Salim
2c2563bc34 fix: gradient checkpointing functools.partial object has no attribute __self__ (#2563) [skip ci] 
* fix: gradient checkpointing causing functools.partial error

* lint

* chore: lint

---------

Co-authored-by: Wing Lian <wing@axolotl.ai>
 2025-04-25 17:02:37 -04:00
Wing Lian
5cb3398460 don't fail on codecov upload for external contributor PRs (#2564) [skip ci] 2025-04-25 15:10:55 -04:00
Dan Saunders
ae1c7ace63 Sequence parallel training context manager (#2553) 
* ctx manager for SP

* updates

* update

* further simplifying

* accommodate both training context managers

* simplifying

* simplifying

* nit

* reorg

* tweak codecov yaml

* add gather post hook, simplify, fixes

* pytest

* pytest fix
 2025-04-25 10:33:54 -04:00
Wing Lian
1447beb132 make sure to validate the config before normalizing so defaults get set (#2554) 
* make sure to validate the config before normalizing so defaults get set

* validation not needed for particular test

* remove duplicate validations

* set qlora correctly
 2025-04-24 13:01:43 -04:00
Dan Saunders
66f41ec6f1 disable codecov pr annotations (#2556) 2025-04-24 08:51:51 -04:00
NanoCode012
85053f4bd4 Fix(doc): add delinearize instruction (#2545) 
* fix: mention to install pytorch before axolotl

* feat(doc): include instruction to delinearize

* fix: update instruction for delinearize with adapter
 2025-04-24 01:03:43 -04:00
Wing Lian
a4d5112ae1 builds for torch 2.7.0 (#2552) 
* builds for torch==2.7.0

* use xformers==0.0.29.post3

* no vllm support with torch 2.7

* update default, fix conditional

* no xformers for 270

* no vllm on 2.7.0 for multigpu test too

* remove deprecated verbose arg from scheduler

* 2.7.0 tests on cpu
 2025-04-24 00:39:31 -04:00
Wing Lian
0d691cc2a7 add base docker image with pytorch 2.7.0 and variant for cuda 12.8 (#2551) 
* add base docker image with pytorch 2.7.0 and variant for cuda 12.8

* my bash is terrible
 2025-04-23 14:59:03 -04:00
Dan Saunders
c4053481ff Codecov fixes / improvements (#2549) 
* adding codecov reporting

* random change

* codecov fixes

* adding missing dependency

* fix

---------

Co-authored-by: Dan Saunders <dan@axolotl.ai>
 2025-04-23 10:33:30 -04:00
NanoCode012
a6d28d19b1 feat: add glm and glm4 multipack and cce (#2546) 
* feat: add glm and glm4 multipack

* feat: add glm4 example

* feat: add cce for glm
 2025-04-23 10:27:51 -04:00
Wing Lian
32e335dd51 fix missing host/port for vllm (#2543) 
* fix missing host/port for vllm

* set tensor parallel size so it doesn't always default to cli override
 2025-04-22 10:16:48 -04:00
Wing Lian
7651550850 make sure to download fixtures for kd test (#2541) 
* make sure to download fixtures for kd test

* use same alpaca dataset
 2025-04-21 10:31:50 -04:00
Wing Lian
341e95aac9 prevent rate limiting to hf when using dispatch batches (#2536) [skip ci] 2025-04-21 10:31:35 -04:00
Catgat
b882dfb63f Fixed Rex Scheduler Warm Up (#2535) [skip ci] 
* Fixed Rex Scheduler Warm Up

* chore: lint

---------

Co-authored-by: Wing Lian <wing@axolotl.ai>
 2025-04-21 10:30:55 -04:00
Wing Lian
b640db1dbc don't run multigpu tests twice, run SP in separate test (#2542) 
* don't run multigpu tests twice, run SP in separate test

* fix multiline
 2025-04-21 10:24:13 -04:00
Chiwan Park
4ce469d32e fix: upgrade liger to 0.5.8 and use native Gemma3 patches (#2527) 
* fix: upgrade liger to 0.5.8 and use native Gemma3 patches

* fix: make lint happy

* doc: update Liger Kernel FLCE support for Gemma 3
 2025-04-18 09:57:40 -07:00
Wing Lian
60a8f0958d zero val fix for beta (#2538) 2025-04-17 17:27:19 -07:00
NanoCode012
9da730d6a4 fix(doc): cut cross entropy installation instructions broken in qmd (#2532) 2025-04-16 15:02:51 -07:00
NanoCode012
32637fad00 fix: preprocess yielding whole dataset to each worker (#2503) [skip ci] 2025-04-16 15:02:35 -07:00