Skip to content

Add vLLM backend for visualqna#905

Merged
poussa merged 1 commit intoopea-project:mainfrom
yongfengdu:issue880
Mar 27, 2025
Merged

Add vLLM backend for visualqna#905
poussa merged 1 commit intoopea-project:mainfrom
yongfengdu:issue880

Conversation

@yongfengdu
Copy link
Copy Markdown
Collaborator

Enable vLLM for lvm-uservice.
Enable vLLM as default LVM backend for VisualQnA.
Update READMEs.
Add PT_HPUGRAPH_DISABLE_TENSOR_CACHE option for vllm, which is necessary to serve LVM models.

Description

The summary of the proposed changes as long as the relevant motivation and context.

Issues

#880
#877

Type of change

List the type of change like below. Please delete options that are not relevant.

  • Bug fix (non-breaking change which fixes an issue)
  • New feature (non-breaking change which adds new functionality)
  • Breaking change (fix or feature that would break existing design and interface)

Dependencies

List the newly introduced 3rd party dependency if exists.

Tests

Helm install with all available values.yaml files.

Enable vLLM for lvm-uservice.
Enable vLLM as default LVM backend for VisualQnA.
Update READMEs.
Add PT_HPUGRAPH_DISABLE_TENSOR_CACHE option for vllm, which is necessary to serve LVM models.

Signed-off-by: Dolpher Du <dolpher.du@intel.com>
@poussa poussa merged commit 485c880 into opea-project:main Mar 27, 2025
19 checks passed
@yongfengdu yongfengdu deleted the issue880 branch April 1, 2025 02:31
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants