Skip to content
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.

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

Update documentation about accelerators #2943

Open
wants to merge 5 commits into
base: main
Choose a base branch
from
Open

Update documentation about accelerators #2943

wants to merge 5 commits into from

Conversation

dtrawins
Copy link
Collaborator

🛠 Summary

CVS-149070
Drop NVIDIA plugin from the maintained list
Drop MULTI plugin - replaced by AUTO with THROUGHPUT_CUMULATIVE hint
Add BATCH virtual device
Refresh documentation about accelerators

🧪 Checklist

  • Unit tests added.
  • The documentation updated.
  • Change follows security best practices.
    ``

@dtrawins dtrawins changed the title Update decumentation about accelerators Update documentation about accelerators Dec 31, 2024

1. Use a .json configuration file to set the `--target_device` parameter with the pattern of: `MULTI:<DEVICE_1>,<DEVICE_2>`.
The order of the devices will define their priority, in this case making `device_1` the primary selection.
OpenVINO Model Server can support using [NPU device](https://docs.openvino.ai/canonical/openvino_docs_install_guides_configurations_for_intel_npu.html)
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

link shows 404

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Comment on lines 56 to 57
Docker image with required dependencies can be build using this procedure:
The docker image of OpenVINO Model Server including support for NVIDIA can be built from sources
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

now it uses one line
image

Suggested change
Docker image with required dependencies can be build using this procedure:
The docker image of OpenVINO Model Server including support for NVIDIA can be built from sources
Docker image with required dependencies can be build using this procedure:
The docker image of OpenVINO Model Server including support for NVIDIA can be built from sources

--model_path /opt/model --model_name resnet --port 9001 \
--plugin_config '{"AUTO_BATCH_TIMEOUT": 200}' \
--target_device BATCH:CPU(16)
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You don't specify what is the (16) in this example - maybe its worth mentioning? Do I understand it correct that it is max batch size when grouping together?

docs/accelerators.md Outdated Show resolved Hide resolved
docs/accelerators.md Outdated Show resolved Hide resolved
char bracket = '(';
auto bracketPos = deviceName.find(bracket);
if (bracketPos != std::string::npos) {
deviceName = deviceName.substr(0, bracketPos);
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why are we doing this? Don't we exclude batching info here?

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

to get physical device and retrieve supported properties, we need to omit content in brackets. For the target device in model compilation, it should be included.

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.

5 participants