chore(deps): update dependency liger-kernel to v0.6.2 #129

red-hat-konflux · 2025-05-03T06:16:01Z

This PR contains the following updates:

Package	Change	Age	Confidence
liger-kernel	`==0.5.4` -> `==0.6.2`

Warning

Some dependencies could not be looked up. Check the warning logs for more information.

Release Notes

linkedin/Liger-Kernel (liger-kernel)

`v0.6.0`: : New Attention Operators, Cosine Similarity Loss, Llama 4, and VLM Patching Updates

Compare Source

Highlights

This release introduces significant improvements to Liger-Kernel, including new operators, support for Llama 4 models, more robust benchmarking automation, and key fixes for patching of vision-language models (VLMs) due to recent transformers refactoring.

Key Changes

New Features & Improvements

Multi-Token Attention by @AndreSlavescu (#689)
Fused Neighborhood Attention by @AndreSlavescu (#732)
Cosine Similarity Loss for Distillation by @Dexterai (#780)
Support for Llama 4 by @Manan17 (#740)
Option to choose fused LCE/CE loss by @connermanuel (#704)
Add block_rms_norm for QK norm by @mdy666 (#731)

Bug Fixes

Vision-language model patching in recent transformers versions (>=4.52.0):
- Qwen2vl, Qwen2_5_vl by @Tcc0403 (#728, #738)
- Llava by @Tcc0403, @Manan17 (#714, #743, #751)
- Gemma3 by @shimizust, @Tcc0403 (#735, #787, #790);
RMS Norm patching by @vaibhavjindal, @BenasdTW (#741, #765)
Hugging Face forward kwargs fix by @llllvvuu (#708)
Fix import tanh by @jue-jue-zi (#762)
Apply monkey patch to instances by @YangKai0616 (#772)

Documentation & CI Fixes

Deploy MkDocs to GitHub Pages by @ParagEkbote (#724)
Robust doc updates by @ParagEkbote (#726, #727)
.idea ignored by @Tcc0403 (#784)
ReadMe, MTA + softmax docs by @AndreSlavescu (#730)
Relax DyT tol, XPU skip MTA by @Tcc0403 (#778)
Paligemma test fixes by @vvvdwbvvv (#785)
Style & test fixes by @Tcc0403, @vaibhavjindal (#736, #794)
Add torchvision for multimodal test by @Tcc0403 (#755)

Benchmarking & Automation

Automated benchmarking and visualization UI in GitHub pages by @Manan17 (#744, #747, #749, #752, #753, #756, #759, #760, #770, #779)

New Contributors

@connermanuel made their first contribution in https://github.com/linkedin/Liger-Kernel/pull/704
@llllvvuu made their first contribution in https://github.com/linkedin/Liger-Kernel/pull/708
@jue-jue-zi made their first contribution in https://github.com/linkedin/Liger-Kernel/pull/762
@YangKai0616 made their first contribution in https://github.com/linkedin/Liger-Kernel/pull/772
@Dexterai made their first contribution in https://github.com/linkedin/Liger-Kernel/pull/780
@vvvdwbvvv made their first contribution in https://github.com/linkedin/Liger-Kernel/pull/785

Full Changelog: linkedin/Liger-Kernel@v0.5.10...v0.6.0

`v0.5.10`: : Qwen3 MOE support, Sparsemax kernel, bug fixes

Compare Source

What's Changed

fix zip bug by @KareemMusleh in https://github.com/linkedin/Liger-Kernel/pull/702
[dpo] set default average_log_prob to False by @cyr0930 in https://github.com/linkedin/Liger-Kernel/pull/693
Rank build status lower by @momochen in https://github.com/linkedin/Liger-Kernel/pull/707
Add support for Qwen3 MoE models by @chiwanpark in https://github.com/linkedin/Liger-Kernel/pull/706
Fix qwen3_moe flaky convergence test by @vaibhavjindal in https://github.com/linkedin/Liger-Kernel/pull/710
Fix empty Medusa head tensors by @chiwanpark in https://github.com/linkedin/Liger-Kernel/pull/698
Sparsemax by @AndreSlavescu in https://github.com/linkedin/Liger-Kernel/pull/687
fix: remove docstring imports in transformer patches by @NanoCode012 in https://github.com/linkedin/Liger-Kernel/pull/712
Increase tests timeout to 45 mins by @vaibhavjindal in https://github.com/linkedin/Liger-Kernel/pull/718
fix modal tests by @shivam15s in https://github.com/linkedin/Liger-Kernel/pull/719
Visualizer Update by @AndreSlavescu in https://github.com/linkedin/Liger-Kernel/pull/717
Sparsemax Documentation by @AndreSlavescu in https://github.com/linkedin/Liger-Kernel/pull/716
element-wise-DyT faster than the origin LigerDyT by @mdy666 in https://github.com/linkedin/Liger-Kernel/pull/673
GRPO Loss kernel fully write by triton, reduce 46G memory by @mdy666 in https://github.com/linkedin/Liger-Kernel/pull/672
Make FLCE compatible with FSDP and PEFT by @astefanutti in https://github.com/linkedin/Liger-Kernel/pull/674
Fix incorrect module patching when using LoRA with modules_to_save by @BenasdTW in https://github.com/linkedin/Liger-Kernel/pull/632
[XPU] Changed how XPU discovery works during setup.py by @Egor-Krivov in https://github.com/linkedin/Liger-Kernel/pull/720
Fix to publish docs on pushes to main branch by @shimizust in https://github.com/linkedin/Liger-Kernel/pull/722
Release 0.5.10 by @shimizust in https://github.com/linkedin/Liger-Kernel/pull/725

New Contributors

@KareemMusleh made their first contribution in https://github.com/linkedin/Liger-Kernel/pull/702
@cyr0930 made their first contribution in https://github.com/linkedin/Liger-Kernel/pull/693
@NanoCode012 made their first contribution in https://github.com/linkedin/Liger-Kernel/pull/712
@mdy666 made their first contribution in https://github.com/linkedin/Liger-Kernel/pull/673
@astefanutti made their first contribution in https://github.com/linkedin/Liger-Kernel/pull/674
@Egor-Krivov made their first contribution in https://github.com/linkedin/Liger-Kernel/pull/720

Full Changelog: linkedin/Liger-Kernel@v0.5.9...v0.5.10

`v0.5.9`: : Adds XPU Setup, GLM-4 & Qwen3 Model Support, Key Bugfixes

Compare Source

What's Changed

update setup.py for installation on xpu by @faaany in https://github.com/linkedin/Liger-Kernel/pull/668
update XPU CI yaml file to use docker container by @faaany in https://github.com/linkedin/Liger-Kernel/pull/669
Add average_log_prob as an init param for LigerFusedLinearDPOLoss by @vaibhavjindal in https://github.com/linkedin/Liger-Kernel/pull/676
add shift label change by @shivam15s in https://github.com/linkedin/Liger-Kernel/pull/683
remove tests that can pass on XPU by @faaany in https://github.com/linkedin/Liger-Kernel/pull/686
Update mkdocs.yml by @shivam15s in https://github.com/linkedin/Liger-Kernel/pull/691
Fix LigerCrossEntropy reduction='none' by @Tcc0403 in https://github.com/linkedin/Liger-Kernel/pull/680
Support GLM-4 models by @intervitens in https://github.com/linkedin/Liger-Kernel/pull/685
Import glm4_lce_forward locally in function by @vaibhavjindal in https://github.com/linkedin/Liger-Kernel/pull/695
Qwen3 model support by @vaibhavjindal in https://github.com/linkedin/Liger-Kernel/pull/692
Use logits_to_keep logic for training runs by @vaibhavjindal in https://github.com/linkedin/Liger-Kernel/pull/696
increase gemma3 multimodal convergence test loss atol by @shivam15s in https://github.com/linkedin/Liger-Kernel/pull/697
Update pyproject.toml by @shivam15s in https://github.com/linkedin/Liger-Kernel/pull/700

New Contributors

@intervitens made their first contribution in https://github.com/linkedin/Liger-Kernel/pull/685

Full Changelog: linkedin/Liger-Kernel@v0.5.8...v0.5.9

`v0.5.8`: : Backward-Compatible Fix

Compare Source

What's Changed

backward compatible initialization by @shivam15s in https://github.com/linkedin/Liger-Kernel/pull/666
Update pyproject.toml by @shivam15s in https://github.com/linkedin/Liger-Kernel/pull/667

Full Changelog: linkedin/Liger-Kernel@v0.5.7...v0.5.8

`v0.5.7`: : Gemma3 Support, XPU Tuning Enhancements, GRPO Improvements, and API Compatibility Fixes

Compare Source

What's Changed

Gemma3 (Text and Multimodal) by @eljandoubi in https://github.com/linkedin/Liger-Kernel/pull/621
Make FLCE compatible with latest XXXForCausalLM.forward() APIs by @Tcc0403 in https://github.com/linkedin/Liger-Kernel/pull/596
do bias addition in tests in float32 to make testing code similar to torch compile by @shivam15s in https://github.com/linkedin/Liger-Kernel/pull/655
[CI] fix siglip dummy config by @yundai424 in https://github.com/linkedin/Liger-Kernel/pull/658
add XPU tuning to JSD by @rmukhopa in https://github.com/linkedin/Liger-Kernel/pull/649
add XPU tuning to Rmsnorm and Layernorm by @Tarakarevu1 in https://github.com/linkedin/Liger-Kernel/pull/653
Fix imports without transformers by @vaibhavjindal in https://github.com/linkedin/Liger-Kernel/pull/659
Use TYPE_CHECKING to fix static-only imports in IDEs etc by @vaibhavjindal in https://github.com/linkedin/Liger-Kernel/pull/660
[kl_div] Modified block and warp sizes for improved performance by @jgtong in https://github.com/linkedin/Liger-Kernel/pull/654
[GRPO] add support for different loss types by @kashif in https://github.com/linkedin/Liger-Kernel/pull/662
Remove unexpected kwargs passing to flce by @Tcc0403 in https://github.com/linkedin/Liger-Kernel/pull/651
reduce number of tests for grpo by @shivam15s in https://github.com/linkedin/Liger-Kernel/pull/663
Update pyproject.toml by @shivam15s in https://github.com/linkedin/Liger-Kernel/pull/665

New Contributors

@rmukhopa made their first contribution in https://github.com/linkedin/Liger-Kernel/pull/649
@Tarakarevu1 made their first contribution in https://github.com/linkedin/Liger-Kernel/pull/653
@jgtong made their first contribution in https://github.com/linkedin/Liger-Kernel/pull/654

Full Changelog: linkedin/Liger-Kernel@v0.5.6...v0.5.7

`v0.5.6`: : Enhancements, Fixes, and Expanded Support (Paligemma, DyT, XPU, Llava, GRPO, and More!)

Compare Source

What's Changed

[JSD] JSD fixes by @kashif in https://github.com/linkedin/Liger-Kernel/pull/609
Paligemma support by @eljandoubi in https://github.com/linkedin/Liger-Kernel/pull/608
Fix hidden size by @eljandoubi in https://github.com/linkedin/Liger-Kernel/pull/612
Add loss_utils for rewriting lce_forward methods by @Tcc0403 in https://github.com/linkedin/Liger-Kernel/pull/614
Update Star History URL by @ryankert01 in https://github.com/linkedin/Liger-Kernel/pull/616
Update README.md by @shivam15s in https://github.com/linkedin/Liger-Kernel/pull/617
language model of paligemma 1 is gemma 1. by @eljandoubi in https://github.com/linkedin/Liger-Kernel/pull/613
Update README to reflect recent changes by @helloworld1 in https://github.com/linkedin/Liger-Kernel/pull/619
Support Dynamic Tanh (DyT) by @Tcc0403 in https://github.com/linkedin/Liger-Kernel/pull/618
Fix incorrect module name when monkey_patch applied to instantiated model by @vaibhavjindal in https://github.com/linkedin/Liger-Kernel/pull/629
[chunked loss] align teacher and student logit shape by @yundai424 in https://github.com/linkedin/Liger-Kernel/pull/634
Fix incorrect condition comment in log_target calculation by @p81sunshine in https://github.com/linkedin/Liger-Kernel/pull/633
Add huggingface llava by @jp1924 in https://github.com/linkedin/Liger-Kernel/pull/524
fix Llava test-bwd failure by @jp1924 in https://github.com/linkedin/Liger-Kernel/pull/639
Fix GRPO to conform with TRL: Fix loss, make tests accurate, correct metrics computation by @shivam15s and @mRSun15 in https://github.com/linkedin/Liger-Kernel/pull/628
add xpu tuning to CE by @mgrabban in https://github.com/linkedin/Liger-Kernel/pull/645
add xpu tuning to FLJSD by @mgrabban in https://github.com/linkedin/Liger-Kernel/pull/647
Change tests to use rocm 6.3 version and tol changes to make liger run on amd by @shivam15s in https://github.com/linkedin/Liger-Kernel/pull/646
Update pyproject.toml by @shivam15s in https://github.com/linkedin/Liger-Kernel/pull/648

New Contributors

@eljandoubi made their first contribution in https://github.com/linkedin/Liger-Kernel/pull/608
@p81sunshine made their first contribution in https://github.com/linkedin/Liger-Kernel/pull/633

Full Changelog: linkedin/Liger-Kernel@v0.5.5...v0.5.6

`v0.5.5`: : Chunk size fixes for JSD; KTO speed fixes; better metrics tests

Compare Source

What's Changed

Infer correct device for AMD HIP device by @helloworld1 in https://github.com/linkedin/Liger-Kernel/pull/587
add out of bounds check to cross entropy by @shivam15s in https://github.com/linkedin/Liger-Kernel/pull/588
Monkeypatch for Qwen2.5-VL by @BenasdTW in https://github.com/linkedin/Liger-Kernel/pull/552
KTO changes to return aux outputs by @vaibhavjindal in https://github.com/linkedin/Liger-Kernel/pull/589
[KTO] Only return summed metrics by @vaibhavjindal in https://github.com/linkedin/Liger-Kernel/pull/591
increase chunk size for distillation and add bias to jsd by @shivam15s in https://github.com/linkedin/Liger-Kernel/pull/590
[CI] Add ROCm 6.3 CI by @tjtanaa in https://github.com/linkedin/Liger-Kernel/pull/506
Fix KTO speed issue by @vaibhavjindal in https://github.com/linkedin/Liger-Kernel/pull/592
Compare means of aggregated outputs in KTO tests by @vaibhavjindal in https://github.com/linkedin/Liger-Kernel/pull/595
Fix means of logps and rewards by @vaibhavjindal in https://github.com/linkedin/Liger-Kernel/pull/597
Add chunk_size param to chunked losses by @RichhLi in https://github.com/linkedin/Liger-Kernel/pull/599
Fix DPO/ORPO typo in readme by @tyler-romero in https://github.com/linkedin/Liger-Kernel/pull/602
version bump by @shivam15s in https://github.com/linkedin/Liger-Kernel/pull/605

New Contributors

@RichhLi made their first contribution in https://github.com/linkedin/Liger-Kernel/pull/599

Full Changelog: linkedin/Liger-Kernel@v0.5.4...v0.5.5

Configuration

📅 Schedule: Branch creation - "after 5am on saturday" (UTC), Automerge - At any time (no schedule defined).

🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.

♻ Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

🔕 Ignore: Close this PR and you won't be reminded about this update again.

If you want to rebase/retry this PR, check this box

To execute skipped test pipelines write comment /ok-to-test.

This PR has been generated by MintMaker (powered by Renovate Bot).

Signed-off-by: red-hat-konflux <126015336+red-hat-konflux[bot]@users.noreply.github.com>

red-hat-konflux bot force-pushed the konflux/mintmaker/main/liger-kernel-0.x branch from bbcdf43 to b797b9c Compare May 10, 2025 07:18

red-hat-konflux bot changed the title ~~chore(deps): update dependency liger-kernel to v0.5.8~~ chore(deps): update dependency liger-kernel to v0.5.9 May 10, 2025

red-hat-konflux bot force-pushed the konflux/mintmaker/main/liger-kernel-0.x branch from b797b9c to db6db39 Compare May 24, 2025 10:13

red-hat-konflux bot changed the title ~~chore(deps): update dependency liger-kernel to v0.5.9~~ Update dependency liger-kernel to v0.5.10 May 24, 2025

red-hat-konflux bot changed the title ~~Update dependency liger-kernel to v0.5.10~~ chore(deps): update dependency liger-kernel to v0.5.10 Jun 21, 2025

red-hat-konflux bot force-pushed the konflux/mintmaker/main/liger-kernel-0.x branch from db6db39 to a0d374f Compare July 12, 2025 08:20

red-hat-konflux bot changed the title ~~chore(deps): update dependency liger-kernel to v0.5.10~~ chore(deps): update dependency liger-kernel to v0.6.0 Jul 12, 2025

red-hat-konflux bot changed the title ~~chore(deps): update dependency liger-kernel to v0.6.0~~ Update dependency liger-kernel to v0.6.0 Jul 19, 2025

red-hat-konflux bot force-pushed the konflux/mintmaker/main/liger-kernel-0.x branch from a0d374f to 4ab8cc4 Compare August 9, 2025 08:55

red-hat-konflux bot changed the title ~~Update dependency liger-kernel to v0.6.0~~ Update dependency liger-kernel to v0.6.1 Aug 9, 2025

Update dependency liger-kernel to v0.6.2

2c4c923

Signed-off-by: red-hat-konflux <126015336+red-hat-konflux[bot]@users.noreply.github.com>

red-hat-konflux bot force-pushed the konflux/mintmaker/main/liger-kernel-0.x branch from 4ab8cc4 to 2c4c923 Compare August 23, 2025 09:20

red-hat-konflux bot changed the title ~~Update dependency liger-kernel to v0.6.1~~ Update dependency liger-kernel to v0.6.2 Aug 23, 2025

red-hat-konflux bot changed the title ~~Update dependency liger-kernel to v0.6.2~~ chore(deps): update dependency liger-kernel to v0.6.2 Aug 30, 2025

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

chore(deps): update dependency liger-kernel to v0.6.2 #129

chore(deps): update dependency liger-kernel to v0.6.2 #129

Uh oh!

red-hat-konflux bot commented May 3, 2025 •

edited

Loading

Uh oh!

Uh oh!

chore(deps): update dependency liger-kernel to v0.6.2 #129

Are you sure you want to change the base?

chore(deps): update dependency liger-kernel to v0.6.2 #129

Uh oh!

Conversation

red-hat-konflux bot commented May 3, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Release Notes

v0.6.2

What's Changed

New Contributors

v0.6.1

What's Changed

New Contributors

v0.6.0: : New Attention Operators, Cosine Similarity Loss, Llama 4, and VLM Patching Updates

Highlights

Key Changes

New Features & Improvements

Bug Fixes

Documentation & CI Fixes

Benchmarking & Automation

New Contributors

v0.5.10: : Qwen3 MOE support, Sparsemax kernel, bug fixes

What's Changed

New Contributors

v0.5.9: : Adds XPU Setup, GLM-4 & Qwen3 Model Support, Key Bugfixes

What's Changed

New Contributors

v0.5.8: : Backward-Compatible Fix

What's Changed

v0.5.7: : Gemma3 Support, XPU Tuning Enhancements, GRPO Improvements, and API Compatibility Fixes

What's Changed

New Contributors

v0.5.6: : Enhancements, Fixes, and Expanded Support (Paligemma, DyT, XPU, Llava, GRPO, and More!)

What's Changed

New Contributors

v0.5.5: : Chunk size fixes for JSD; KTO speed fixes; better metrics tests

What's Changed

New Contributors

Configuration

Uh oh!

Uh oh!

red-hat-konflux bot commented May 3, 2025 •

edited

Loading

`v0.6.2`

`v0.6.1`

`v0.6.0`: : New Attention Operators, Cosine Similarity Loss, Llama 4, and VLM Patching Updates

`v0.5.10`: : Qwen3 MOE support, Sparsemax kernel, bug fixes

`v0.5.9`: : Adds XPU Setup, GLM-4 & Qwen3 Model Support, Key Bugfixes

`v0.5.8`: : Backward-Compatible Fix

`v0.5.7`: : Gemma3 Support, XPU Tuning Enhancements, GRPO Improvements, and API Compatibility Fixes

`v0.5.6`: : Enhancements, Fixes, and Expanded Support (Paligemma, DyT, XPU, Llava, GRPO, and More!)

`v0.5.5`: : Chunk size fixes for JSD; KTO speed fixes; better metrics tests