Commit Graph

2691 Commits

Author SHA1 Message Date
Zu Kim
460346ed13 Add a support for label annotations (image/label/string and image/label/confidence). Also fixed some clang tidy issues.
PiperOrigin-RevId: 553900667
2023-08-04 13:43:44 -07:00
MediaPipe Team
11508f2291 Internal Change
PiperOrigin-RevId: 553652444
2023-08-03 18:49:35 -07:00
MediaPipe Team
360959e325 Replace some size EXPECTs by ASSERTs
PiperOrigin-RevId: 553555650
2023-08-03 12:27:52 -07:00
MediaPipe Team
a0b91e4062 Add a GpuOrigin parameter to TensorConverterCalculator
The parameter superseeds flip_vertically. GpuOrigin works more generally than flip_vertically because CONVENTIONAL works on both iOS (no flip) and Android (yes flip). If not set, the calculator falls back to flip_vertically for backwards compatibility.

Note that web demos actually use TOP_LEFT image orientation, so they shouldn't be flipped, but they still are by CONVENTIONAL. That's being discussed right now.

PiperOrigin-RevId: 553400525
2023-08-03 01:40:13 -07:00
MediaPipe Team
9325af0af3 vlog default executor and its config usage
PiperOrigin-RevId: 553298440
2023-08-02 17:01:05 -07:00
MediaPipe Team
e56636b6d1 internal change
PiperOrigin-RevId: 553250547
2023-08-02 14:06:16 -07:00
MediaPipe Team
366a3290cf Change to add the w_avg latent code to style encoding before layer swapping. This is a bug in the previous code. Also set training=True for encoder since this affect the encoding performance.
PiperOrigin-RevId: 553234376
2023-08-02 13:11:22 -07:00
MediaPipe Team
6e54d8c204 Log stack traces for combined CalculatorGraph statuses
PiperOrigin-RevId: 553111356
2023-08-02 05:10:52 -07:00
MediaPipe Team
557ed0b1ea Add tensorflow-addons to model_maker requirements.txt
PiperOrigin-RevId: 552610011
2023-07-31 15:36:25 -07:00
MediaPipe Team
6f916a001c Fix crash in SavePngTestOutput
Do not call SavePngTestOutput in CompareAndSaveImageOutput in case diff_img is null. This can happen if for instance the expected and the actual image have non-matching format or size. Currently, this crashes.
Support single channel golden images.

PiperOrigin-RevId: 552519834
2023-07-31 10:20:06 -07:00
Sebastian Schmidt
d392f8ad98 Ensure that -std=c++14/17 is the first argument passed to Glog
PiperOrigin-RevId: 552509553
2023-07-31 09:47:32 -07:00
Sebastian Schmidt
81cf7fa173 Updat WASM binaries for 0.10.3 release
PiperOrigin-RevId: 551975834
2023-07-28 16:17:30 -07:00
MediaPipe Team
8e313b4b0c Fix typo in model maker requirements.txt
PiperOrigin-RevId: 551973577
2023-07-28 16:07:12 -07:00
Sebastian Schmidt
b4bcfab4f5 Remove extra letter from text classifier API
PiperOrigin-RevId: 551942087
2023-07-28 13:56:56 -07:00
Sebastian Schmidt
8ab9185c1d Use C+++ 17 for Glog only on Windows
PiperOrigin-RevId: 551928369
2023-07-28 12:58:42 -07:00
MediaPipe Team
3f7752561b No public description
PiperOrigin-RevId: 551914786
2023-07-28 12:05:38 -07:00
MediaPipe Team
9edb059d9f No public description
PiperOrigin-RevId: 551868738
2023-07-28 09:09:39 -07:00
MediaPipe Team
7db0c1944b Internal change
PiperOrigin-RevId: 551789915
2023-07-28 02:29:52 -07:00
MediaPipe Team
db9a72a5df Internal Changes
PiperOrigin-RevId: 551674542
2023-07-27 16:35:35 -07:00
MediaPipe Team
5c007558f8 internal change.
PiperOrigin-RevId: 551645248
2023-07-27 14:45:18 -07:00
Sebastian Schmidt
4d5c6bd33a Internal
PiperOrigin-RevId: 551625147
2023-07-27 13:33:14 -07:00
Sebastian Schmidt
fdea10d230 Add C Headers for Text Classifier
PiperOrigin-RevId: 551618765
2023-07-27 13:09:49 -07:00
MediaPipe Team
5b31f1e3e9 Update glog to latest commit
PiperOrigin-RevId: 551601991
2023-07-27 12:07:20 -07:00
Sebastian Schmidt
7d9cb4ee67 No public description
PiperOrigin-RevId: 551586945
2023-07-27 11:17:54 -07:00
MediaPipe Team
f3f9e71ccb No public description
PiperOrigin-RevId: 551549511
2023-07-27 09:15:46 -07:00
MediaPipe Team
6de275834d internal change.
PiperOrigin-RevId: 551366789
2023-07-26 17:59:06 -07:00
Sebastian Schmidt
c9d79a0076 Rollback of "Fix duplicate condition error in :resource_util"
PiperOrigin-RevId: 551332734
2023-07-26 15:31:51 -07:00
Sebastian Schmidt
dad46e1e90 Update glog to 0.6
PiperOrigin-RevId: 551330044
2023-07-26 15:20:34 -07:00
Sebastian Schmidt
f156397e8f Fix Android build with any Protos
PiperOrigin-RevId: 551325541
2023-07-26 15:04:38 -07:00
MediaPipe Team
fa5c1b03d2 No public description
PiperOrigin-RevId: 551277242
2023-07-26 12:08:33 -07:00
Sebastian Schmidt
87b925795d Update glog to 0.6
PiperOrigin-RevId: 551269455
2023-07-26 11:41:24 -07:00
Sebastian Schmidt
750f498b14 Internal
PiperOrigin-RevId: 551247471
2023-07-26 10:33:19 -07:00
Prianka Liz Kariat
d2a86341bd Added refactored iOS vision task runner sources 2023-07-26 20:02:00 +05:30
MediaPipe Team
1f6851c577 C++ Image segmenter add output size parameters.
PiperOrigin-RevId: 550995124
2023-07-25 14:22:11 -07:00
MediaPipe Team
bd7888cc0c 1. Move evaluation onto GPU/TPU hardware if available.
2. Move desired_precision and desired_recall from evaluate to hyperparameters so recall@precision metrics will be reported for both training and evaluation. This also fixes a bug where recompiling the model with the previously initialized metric objects would not properly reset the metric states.
3. Remove redundant label_names from create_... class methods in text_classifier. This information is already provided by the datasets.
4. Change loss function to FocalLoss.
5. Re-enable text_classifier unit tests using ExBert
6. Add input names to avoid flaky auto-assigned input names.

PiperOrigin-RevId: 550992146
2023-07-25 14:12:26 -07:00
MediaPipe Team
85c3fed70a Add class weights to core hyperparameters and classifier library.
PiperOrigin-RevId: 550962843
2023-07-25 12:29:46 -07:00
MediaPipe Team
62538a9496 No public description
PiperOrigin-RevId: 550954023
2023-07-25 11:57:08 -07:00
MediaPipe Team
113c9b30c2 No public description
PiperOrigin-RevId: 550616150
2023-07-24 11:09:31 -07:00
Copybara-Service
66cceddb5e Merge pull request #4639 from priankakariatyml:ios-image-segmenter-container-utils
PiperOrigin-RevId: 550608973
2023-07-24 10:46:56 -07:00
Prianka Liz Kariat
72c62f7d5d Added iOS Image Segmenter Header 2023-07-24 20:38:16 +05:30
MediaPipe Team
25b01784de Fix documentation
PiperOrigin-RevId: 549968822
2023-07-21 09:33:48 -07:00
MediaPipe Team
9af637b125 Java API add visibility and presence for landmarks.
PiperOrigin-RevId: 549709256
2023-07-20 12:42:01 -07:00
Copybara-Service
236a36e39a Merge pull request #4629 from priankakariatyml:ios-vision-library-fixes
PiperOrigin-RevId: 549659874
2023-07-20 09:52:11 -07:00
Prianka Liz Kariat
540f4f7fe6 Fixed swift name of iOS face landmarker delegate 2023-07-20 15:57:37 +05:30
Prianka Liz Kariat
3198ccf6a5 Added missing headers in ios vision framework build 2023-07-20 15:57:16 +05:30
Steven Hickson
e47af74b15 Adding support for 2 things in tensors_to_image_calculator:
1) 1 channel support for conversion after inference.
2) multitask support by allowing for different tensor outputs.

PiperOrigin-RevId: 549412331
2023-07-19 13:41:46 -07:00
MediaPipe Team
085840388b Move waitOnCpu and waitOnGpu out of the synchronized block, which can cause deadlock.
PiperOrigin-RevId: 549217916
2023-07-18 23:42:01 -07:00
MediaPipe Team
4e72fcf0cb Replace CHECK with RET_CHECK in GetContract() implementation from six calculators.
PiperOrigin-RevId: 549158984
2023-07-18 17:38:44 -07:00
MediaPipe Team
4c60fe7365 add pose landmarks connections in C++ API
PiperOrigin-RevId: 549108310
2023-07-18 14:21:02 -07:00
MediaPipe Team
9b00582f21 add hand landmarks connections in C++ API.
PiperOrigin-RevId: 549108307
2023-07-18 14:16:27 -07:00