Skip to content
This repository was archived by the owner on Sep 4, 2025. It is now read-only.

Sync release to main for rhoai-2.14 #172

Merged
merged 420 commits into from
Sep 30, 2024
Merged

Sync release to main for rhoai-2.14 #172

merged 420 commits into from
Sep 30, 2024

Conversation

@vaibhavjainwiz
Copy link
Member

@vaibhavjainwiz vaibhavjainwiz commented Sep 27, 2024 •
edited by rpancham
Loading

LiuXiaoxuanPKU and others added 30 commits September 11, 2024 14:07
@LiuXiaoxuanPKU @youkaichao
[Speculative Decoding] Test refactor (vllm-project#8317)
775f00f 
Co-authored-by: youkaichao <[email protected]>
@patrickvonplaten @ywang96
Pixtral (vllm-project#8377)
d394787 
Co-authored-by: Roger Wang <[email protected]>
@simon-mo
Bump version to v0.6.1 (vllm-project#8379)
3fd2b0d
@comaniac
[MISC] Dump model runner inputs when crashing (vllm-project#8305)
a65cb16
@youkaichao
[misc] remove engine_use_ray (vllm-project#8126)
f842a7a
@WoosukKwon
[TPU] Use Ray for default distributed backend (vllm-project#8389)
b71c956
@mgoin
Fix the AMD weight loading tests (vllm-project#8390)
b6c75e1
@tomeras91
[Bugfix]: Fix the logic for deciding if tool parsing is used (vllm-pr…
5a60699 
…oject#8366)
@blueyo0
[Gemma2] add bitsandbytes support for Gemma2 (vllm-project#8338)
1bf2dd9
@kevin314
[Misc] Raise error when using encoder/decoder model with cpu backend (v…
295c473 
…llm-project#8355)
@WoosukKwon
[Misc] Use RoPE cache for MRoPE (vllm-project#8396)
42ffba1
@youkaichao
[torch.compile] hide slicing under custom op for inductor (vllm-proje…
7de49aa 
…ct#8384)
@ywang96
[Hotfix][VLM] Fixing max position embeddings for Pixtral (vllm-projec…
520ca38 
…t#8399)
@Isotr0py @DarkLight1337
[Bugfix] Fix InternVL2 inference with various num_patches (vllm-proje…
e56bf27 
…ct#8375)

Co-authored-by: DarkLight1337 <[email protected]>
@alex-jw-brooks @DarkLight1337
[Model] Support multiple images for qwen-vl (vllm-project#8247)
c6202da 
Signed-off-by: Alex-Brooks <[email protected]>
Co-authored-by: Cyrus Leung <[email protected]>
Co-authored-by: DarkLight1337 <[email protected]>
@lnykww
[BugFix] lazy init _copy_stream to avoid torch init wrong gpu instance (
8a23e93 
vllm-project#8403)
@vegaluisjose
[BugFix] Fix Duplicate Assignment in Hermes2ProToolParser (vllm-proje…
1f0c75a 
…ct#8423)
@joerunde
[Bugfix] Offline mode fix (vllm-project#8376)
f2e263b 
Signed-off-by: Joe Runde <[email protected]>
@SolitaryThinker
[multi-step] add flashinfer backend (vllm-project#7928)
a6c0f36
@njhill
[Core] Add engine option to return only deltas or final output (vllm-…
551ce01 
…project#7381)
@alexm-redhat
[Bugfix] multi-step + flashinfer: ensure cuda graph compatible (vllm-…
0198772 
…project#8427)
@ywang96
[Hotfix][Core][VLM] Disable chunked prefill by default and prefix cac…
c163694 
…hing for multimodal models (vllm-project#8425)
@ywang96
[CI/Build] Disable multi-node test for InternVL2 (vllm-project#8428)
b61bd98
@patrickvonplaten
[Hotfix][Pixtral] Fix multiple images bugs (vllm-project#8415)
d31174a
@wenxcs
[Bugfix] Fix weight loading issue by rename variable. (vllm-project#8293
a480939 
)
@ywang96
[Misc] Update Pixtral example (vllm-project#8431)
360ddbd
@dsikka
[BugFix] fix group_topk (vllm-project#8430)
8f44a92
@DarkLight1337
[Core] Factor out input preprocessing to a separate class (vllm-proje…
5ec9c0f 
…ct#7329)
@ShangmingCai
[Bugfix] Mapping physical device indices for e2e test utils (vllm-pro…
40c3965 
…ject#8290)
@DarkLight1337
[Bugfix] Bump fastapi and pydantic version (vllm-project#8435)
3f79bc3
dtrifiro and others added 15 commits September 27, 2024 17:14
@dtrifiro
Dockerfile.ubi: bump python to 3.12
b30f9ed
@dtrifiro
Dockerfile.ubi: bump flashinfer to 0.1.6
8238489
@dtrifiro
Dockerfile.rocm.ubi: do not use nightly pytorch_triton
67080c2
@dtrifiro
Dockerfile.ubi: fix PYTHON_VERSION arg usage
010c1bd 
fixup to f5387d0 (openshift/release#56770)
@dtrifiro
Dockerfile.rocm.ubi: move microdnf update in base stage
f91432f
@dtrifiro
Dockerfile.rocm.ubi: bump torch version to 2.5.0.dev20240912+rocm6.1
b1b179f
@dtrifiro
Dockerfile.rocm.ubi: get rid of build triton stage
cd4b748 
this is a torch dependency when installed from the pytorch/rocm6.1
index: https://download.pytorch.org/whl/nightly/rocm6.1
@dtrifiro
Dockerfile.rocm.ubi: add setuptools-scm build dependency
38247ba
@dtrifiro
Dockerfile.ubi: add VLLM_FA_CMAKE_GPU_ARCHES
451470c
@dtrifiro
Dockerfile.ubi: use COPY . . to make repo available when building wheel
81a8400
@stevegrubb @dtrifiro
This sets the vllm build to a Release build type, builds libsodium,
ec1f663 
fixes bad permissions on a logging directory, installs libsodium,
and adds a LD_LIBRARY_PATH to fixup python bindings for 2 packages.
@stevegrubb @dtrifiro
Move libsodium install
d16bf47 
This moves the install up a couple lines.
@stevegrubb @dtrifiro
Correct logging directory permissions
f63fbdd 
The /var/log/rocm_smi_lib/ directory was world writable. It
is fixed now so that it is world readable. Similarly,
the file in it's directory is also world writable, it is
now world readable.
@NickLucche @dtrifiro
bump tgis adapter to v0.5.0
902985d
@vaibhavjainwiz
Merge branch 'main' into sync_release
9d9bb9c
@openshift-ci openshift-ci bot requested review from heyselbi and z103cb September 27, 2024 17:03
@openshift-ci
Copy link

openshift-ci bot commented Sep 27, 2024

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: vaibhavjainwiz

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

@rpancham
Copy link

rpancham commented Sep 27, 2024

/lgtm

@openshift-ci openshift-ci bot added the lgtm label Sep 27, 2024
@openshift-ci
Copy link

openshift-ci bot commented Sep 27, 2024 •
edited
Loading

@vaibhavjainwiz: The following test failed, say /retest to rerun all failed tests or /retest-required to rerun all mandatory failed tests:

Test name Commit Details Required Rerun command
ci/prow/smoke-test 9d9bb9c link true /test smoke-test

Full PR test history. Your PR dashboard.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. I understand the commands that are listed here.

@vaibhavjainwiz
Copy link
Member Author

vaibhavjainwiz commented Sep 28, 2024

/test images

@vaibhavjainwiz
Copy link
Member Author

vaibhavjainwiz commented Sep 28, 2024

/test pr-image-mirror

@vaibhavjainwiz vaibhavjainwiz merged commit 2f1734e into opendatahub-io:release Sep 30, 2024
2 of 3 checks passed
@vaibhavjainwiz vaibhavjainwiz deleted the sync_release branch October 4, 2024 10:24
prarit pushed a commit to prarit/vllm that referenced this pull request Oct 21, 2024
@divakar-amd
[fix] moe padding for reading correct tuned config (opendatahub-io#172)
b3fc9f4
fialhocoelho added a commit that referenced this pull request Oct 24, 2024
@fialhocoelho
using adapter version merged with PR #172
080973c 
Signed-off-by: Jefferson Fialho <[email protected]>
riprasad referenced this pull request in red-hat-data-services/vllm Oct 25, 2024
@kzawora-intel
1.17 documentation update (#172)
6f047d8
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Reviewers

@heyselbi heyselbi Awaiting requested review from heyselbi

@z103cb z103cb Awaiting requested review from z103cb

Assignees

@rpancham rpancham

Labels

approved lgtm

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.