-
-
Notifications
You must be signed in to change notification settings - Fork 1.1k
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
example: PyTorch Live integration #1485
Draft
mrousavy
wants to merge
55
commits into
main
Choose a base branch
from
example/pytorch
base: main
Could not load branches
Branch not found: {{ refName }}
Loading
Could not load tags
Nothing to show
Loading
Are you sure you want to change the base?
Some commits from the old base branch may be removed from the timeline,
and old review comments may become outdated.
Draft
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
* chore: Upgrade Example to RN 0.71 * chore: Upgrade all libs * fix: Fix CameraRoll installation * Update Gradle Tools * fix: Fix buildscripts * Clean out build.gradle * fix: Fix Kotlin setup * fix: Move kotlin-android dependency to lib * Move `_setGlobalConsole` * Update gradle-wrapper.properties * Rebuild lockfiles * chore: Update build:gradle * Update StatusBarBlurBackground.tsx * Use Java 11 in Workflows * Update MediaPage.tsx * Add `google` repository to build.gradle * Double Java Heap size * Increase heap size * Alternative args * Update build.gradle
* Setup RN Worklets * Use RN Worklets on iOS * Fix console * Add `installFrameProcessorBindings()` function * Add `FrameProcessorPlugins` proxy (BREAKING CHANGE) * Clean up docs * Update FRAME_PROCESSORS.mdx * Use RN Worklets 0.2.5 * feat: Android build setup * Rewrite Android Frame Processor Part * Update CMakeLists.txt * fix: Add react-native-worklets Gradle dependency * Update Podfile.lock * fix build * gradle:7.4.1 * Init JSI Bindings in method on Android * Fix Folly flags * fix: Init `FrameProcessorRuntimeManager` later * fix: Wrap in `<GestureHandlerRootView>` * Refactor plugins * fix: Remove enableFrameProcessors * Install RN Worklets from current GH master * Update babel.config.js * Update CameraViewModule.kt * Update ImageProxyUtils.java * feat: Upgrade to Reanimated v3 * fix: Fix crash on Worklet init * Update RN Worklets to latest master * fix: Simplify FP Plugins Proxy
…1472) Before, Frame Processors ran on a separate Thread. After, Frame Processors run fully synchronous and always at the same FPS as the Camera. Two new functions have been introduced: * `runAtTargetFps(fps: number, func: () => void)`: Runs the given code as often as the given `fps`, effectively throttling it's calls. * `runAsync(frame: Frame, func: () => void)`: Runs the given function on a separate Thread for Frame Processing. A strong reference to the Frame is held as long as the function takes to execute. You can use `runAtTargetFps` to throttle calls to a specific API (e.g. if your Camera is running at 60 FPS, but you only want to run face detection at ~25 FPS, use `runAtTargetFps(25, ...)`.) You can use `runAsync` to run a heavy algorithm asynchronous, so that the Camera is not blocked while your algorithm runs. This is useful if your main sync processor draws something, and your async processor is doing some image analysis on the side. You can also combine both functions. Examples: ```js const frameProcessor = useFrameProcessor((frame) => { 'worklet' console.log("I'm running at 60 FPS!") }, []) ``` ```js const frameProcessor = useFrameProcessor((frame) => { 'worklet' console.log("I'm running at 60 FPS!") runAtTargetFps(10, () => { 'worklet' console.log("I'm running at 10 FPS!") }) }, []) ``` ```js const frameProcessor = useFrameProcessor((frame) => { 'worklet' console.log("I'm running at 60 FPS!") runAsync(frame, () => { 'worklet' console.log("I'm running on another Thread, I can block for longer!") }) }, []) ``` ```js const frameProcessor = useFrameProcessor((frame) => { 'worklet' console.log("I'm running at 60 FPS!") runAtTargetFps(10, () => { 'worklet' runAsync(frame, () => { 'worklet' console.log("I'm running on another Thread at 10 FPS, I can block for longer!") }) }) }, []) ```
* fix: Fix CI for "Build Android" * update versions * Update Gemfile.lock * format swift * fix: Fix swift lint * Update .swiftlint.yml * Use C++17 for lint * fix: Fix C++ lints
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What
Uses PyTorch Core to convert a Frame to a Image. This would allow us to run any ML Model inside VisionCamera straight from JS - no native code (Frame Processor Plugins) - proof of concept PR: facebookresearch/playtorch#199
Changes
Tested on
Related issues