Name: TensorFlow Lite Support
Short Name: tflite-support
URL: https://github.com/tensorflow/tflite-support
Version: d3a8a112e5d3f83696e665f76f660f2b5dd47052
Revision: d3a8a112e5d3f83696e665f76f660f2b5dd47052
Date: 2023-09-27
License: Apache 2.0
License File: LICENSE
Security Critical: Yes
Shipped: yes
CPEPrefix: unknown
Description:
TFLite Support is a toolkit that helps users to develop ML and deploy TFLite
models onto mobile devices. It works cross-Platform and is supported on
Java, C++ (WIP), and Swift (WIP).
Third party dependencies:
- tflite
- libzip
- utf
- tensorflow-text
Local modifications with their associated descriptions are contained in the
patches/ subdirectory.
Update process:
1) First, update Chrome's copy from the main tflite support repository:
```
pushd third_party/tflite_support/
rm -rf src/
git clone https://github.com/tensorflow/tflite-support/ src
rm -rf src/.git/
git add src
popd
git commit -m "Initial roll."
```
2) Apply the patches in patches/:
```
git am --keep-cr third_party/tflite_support/patches/*
```
3) This works somewhat like a rebase, and conflicts may need to be resolved
along the way.
4) If additional patches are needed, record each distinct patch as its own git
commit. A reasonably-sized patch would be "delete component <X> which is not
used in Chrome". A patch named "fix the build for Chrome" would probably be
too generic and large.
5) Regenerate the patches/ subdirectory:
```
git rm third_party/tflite_support/patches/*
git format-patch HEAD -<number of changes> -o third_party/tflite_support/patches
git add third_party/tflite_support/patches/*
```
Important: do not proceed to the next steps until all local patches have been
updated and applied!
6) Remove the testdata directory, which contains many large binary artifacts:
```
git rm -r third_party/tflite_support/src/tensorflow_lite_support/cc/test/testdata
```
This is not tracked as a patchfile, as `git format-patch` can generate very
large diffs for this.