Project import generated by Copybara.

GitOrigin-RevId: 50714fe28298d7b707eff7304547d89d6ec34a54

mediapipe/examples/desktop/README.md

@@ -9,7 +9,9 @@ bazel build -c opt mediapipe/examples/desktop/hello_world:hello_world
 and then run it using:
 
 ```
-bazel-bin/mediapipe/examples/desktop/hello_world/hello_world --logtostderr
+export GLOG_logtostderr=1
+
+bazel-bin/mediapipe/examples/desktop/hello_world/hello_world
 ```
 
 **TFlite Object Detection**
@@ -23,10 +25,11 @@ bazel build -c opt mediapipe/examples/desktop/object_detection:object_detection_
 and run it using:
 
 ```
+export GLOG_logtostderr=1
+
 bazel-bin/mediapipe/examples/desktop/object_detection/object_detection_tflite \
   --calculator_graph_config_file=mediapipe/graphs/object_detection/object_detection_desktop_tflite_graph.pbtxt \
-  --input_side_packets=input_video_path=/path/to/input/file,output_video_path=/path/to/output/file \
-  --alsologtostderr
+  --input_side_packets=input_video_path=/path/to/input/file,output_video_path=/path/to/output/file
 ```
 
 **TensorFlow Object Detection**
@@ -34,6 +37,8 @@ bazel-bin/mediapipe/examples/desktop/object_detection/object_detection_tflite \
 To build the object detection demo using a TensorFlow model on desktop, use:
 
 ```
+export GLOG_logtostderr=1
+
 bazel build -c opt mediapipe/examples/desktop/object_detection:object_detection_tensorflow \
   --define MEDIAPIPE_DISABLE_GPU=1
 ```
@@ -41,8 +46,68 @@ bazel build -c opt mediapipe/examples/desktop/object_detection:object_detection_
 and run it using:
 
 ```
+export GLOG_logtostderr=1
+
 bazel-bin/mediapipe/examples/desktop/object_detection/object_detection_tensorflow  \
   --calculator_graph_config_file=mediapipe/graphs/object_detection/object_detection_desktop_tensorflow_graph.pbtxt  \
   --input_side_packets=input_video_path=/path/to/input/file,output_video_path=/path/to/output/file
-  --alsologtostderr
 ```
+
+**TFlite Hand Detection**
+
+To build the hand detection demo using a TFLite model on desktop, use:
+
+```
+bazel build -c opt mediapipe/examples/desktop/hand_tracking:hand_tracking_tflite --define MEDIAPIPE_DISABLE_GPU=1
+```
+
+and run it using:
+
+```
+export GLOG_logtostderr=1
+
+bazel-bin/mediapipe/examples/desktop/hand_tracking/hand_tracking_tflite \
+  --calculator_graph_config_file=mediapipe/graphs/hand_tracking/hand_detection_desktop.pbtxt \
+  --input_side_packets=input_video_path=/path/to/input/file,output_video_path=/path/to/output/file
+```
+
+**TFlite Hand Tracking**
+
+To build the hand tracking demo using a TFLite model on desktop, use:
+
+```
+bazel build -c opt mediapipe/examples/desktop/hand_tracking:hand_tracking_tflite --define MEDIAPIPE_DISABLE_GPU=1
+```
+
+and run it using:
+
+```
+export GLOG_logtostderr=1
+
+bazel-bin/mediapipe/examples/desktop/hand_tracking/hand_tracking_tflite \
+  --calculator_graph_config_file=mediapipe/graphs/hand_tracking/hand_tracking_desktop.pbtxt \
+  --input_side_packets=input_video_path=/path/to/input/file,output_video_path=/path/to/output/file
+```
+
+**TFlite Multi-Hand Tracking**
+
+To build the multi-hand tracking demo using a TFLite model on desktop, use:
+
+```
+bazel build -c opt mediapipe/examples/desktop/multi_hand_tracking:multi_hand_tracking_tflite --define MEDIAPIPE_DISABLE_GPU=1
+```
+
+and run it using:
+
+```
+export GLOG_logtostderr=1
+
+bazel-bin/mediapipe/examples/desktop/multi_hand_tracking/multi_hand_tracking_tflite \
+  --calculator_graph_config_file=mediapipe/graphs/hand_tracking/multi_hand_tracking_desktop.pbtxt \
+  --input_side_packets=input_video_path=/path/to/input/file,output_video_path=/path/to/output/file
+```
+
+To change the number of hands to `x` in this application, change:
+
+1. `min_size:x` in `CollectionHasMinSizeCalculatorOptions` in `mediapipe/graphs/hand_tracking/multi_hand_tracking_desktop.pbtxt`.
+2. `max_vec_size:x` in `ClipVectorSizeCalculatorOptions` in `mediapipe/examples/dekstop/hand_tracking/subgraphs/multi_hand_detection_cpu.pbtxt`.

mediapipe/examples/desktop/media_sequence/kinetics_dataset.py

@@ -93,15 +93,15 @@ FILEPATTERN = "kinetics_700_%s_25fps_rgb_flow"
 SPLITS = {
     "train": {
         "shards": 1000,
-        "examples": 541632
+        "examples": 541490
     },
     "validate": {
         "shards": 100,
-        "examples": 34727
+        "examples": 34715
     },
     "test": {
         "shards": 100,
-        "examples": 69347
+        "examples": 69321
     },
     "custom": {
         "csv": None,  # Add a CSV for your own data here.