pytorch / pytorch/executorch

Question about programmatically running inference on Android-based custom OS with vulkan delegate

Open
#10,602 10 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

module: vulkan
Dominant language
Python
Stars
5k
Forks
1.2k
Avg merge
2d 10h
Merged PRs (30d)
581

Description

Hi team,

I have built the vulkan delegate for android by referring to the docs - https://pytorch.org/executorch/stable/native-delegates-executorch-vulkan-delegate.html

I am running a custom Linux OS which is based on Android's Linux Kernel. I am able to use the above binary to run models on my GPU like -

vulkan_executor_runner –-model_path .pte

How can I do this programmatically so that I can obtain the outputs for post-processing? I found this example for C++ but how do I force the implementation to use my "vulkan_executor_runner"

cc @SS-JIA @manuelcandales @cbilgin

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start with the linked Vulkan delegate documentation and the C++ example mentioned in the issue. Determine how programmatic inference can obtain outputs on the Android-based custom OS and whether the existing vulkan_executor_runner binary can be used directly. Done means documenting a supported integration path or clearly stating the limitation.

Written by the indexing model from the issue text.

Assessment

Tech stack
android, cpp, machine-learning
Domain
api, machine-learning, mobile
Issue type
Documentation
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.