You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: docs/demultiplexing.md
+2-2Lines changed: 2 additions & 2 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -1,7 +1,7 @@
1
1
# Demultiplexing in Directed Acyclic Graph Scheduler
2
2
3
3
## Pipeline branching
4
-
Directed Acyclic Graph Scheduler allows creating pipelines with optional parameter `demultiply_count: `N` which adds ability to any node to slice outputs into `N` separate sub outputs and branch pipeline execution into `N` sub pipelines. Following nodes will be executed by event loop `N` times independently from each other and results will be gathered and packed into one output just before sending a response. Additionally `gather_from_node: <node_name>` parameter can be used to specify gathering at any point in Directed Acyclic Graph.`demultiply_count` node parameter value has to match first dimension of all node outputs. All node outputs have to have at least 2 dimensions.
4
+
Directed Acyclic Graph Scheduler allows creating pipelines with optional parameter `demultiply_count: N` which adds ability to any node to slice outputs into `N` separate sub outputs and branch pipeline execution into `N` sub pipelines. Following nodes will be executed by event loop `N` times independently from each other and results will be gathered and packed into one output just before sending a response. Additionally `gather_from_node: <node_name>` parameter can be used to specify gathering at any point in Directed Acyclic Graph.`demultiply_count` node parameter value has to match first dimension of all node outputs. All node outputs have to have at least 2 dimensions.
5
5
6
6
## Basic demultiplexer example and metadata explanation
7
7
This example contains 2 consecutive models:
@@ -12,7 +12,7 @@ This example has no `gather_from_node` parameter specified in config.json config
12
12
13
13

14
14
15
-
Example configuration file with one demultiplexer:
15
+
Example configuration file with one demultiplexer (remove everything after arrows in lines containing those):
Copy file name to clipboardExpand all lines: docs/developer_guide.md
+7-5Lines changed: 7 additions & 5 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -27,8 +27,10 @@ In-case of problems, see <a href="#debug">Debugging</a>.
27
27
### Step 1: Compile source code
28
28
1. Build the development `openvino/model_server-build` Docker* image
29
29
```bash
30
-
make docker_build
30
+
make docker_build
31
+
```
31
32
or
33
+
```
32
34
make docker_build DLDT_PACKAGE_URL=<URL>
33
35
```
34
36
> **Note**: URL to OpenVINO Toolkit package can be received after registration on [OpenVINO™ Toolkit website](https://software.intel.com/en-us/openvino-toolkit/choose-download)
@@ -63,8 +65,8 @@ In-case of problems, see <a href="#debug">Debugging</a>.
63
65
* With a Docker cache :
64
66
65
67
```bash
66
-
OVMS_CPP_DOCKER_IMAGE=<unique_image_name> make docker_build
67
-
OVMS_CPP_DOCKER_IMAGE=<unique_image_name> make test_functional
68
+
OVMS_CPP_DOCKER_IMAGE=<replace_with_unique_image_name> make docker_build
69
+
OVMS_CPP_DOCKER_IMAGE=<replace_with_unique_image_name> make test_functional
68
70
OVMS_CPP_CONTAINTER_PORT=<unique_network_port> make test_perf
69
71
```
70
72
@@ -265,10 +267,10 @@ Two debugging options are available. Click on the required option :
Copy file name to clipboardExpand all lines: docs/east_ocr.md
+2-2Lines changed: 2 additions & 2 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -12,7 +12,7 @@ Below is depicted the graph implementing a complete OCR pipelines.
12
12
13
13

14
14
15
-
It includes the following Nodes:
15
+
It includes the following nodes:
16
16
- Model east-resnet50 - inference execution which takes the user image as input. It returns two outputs including information about all detected boxes, their location and scores.
17
17
- Custom node east_ocr - it includes C++ implementation of east-resnet50 model results processing. It analyses the detected boxes coordinates, filters the results
18
18
based on the configurable score level threshold and and applies non-max suppression algorithm to remove overlaping boxes. Finally the custom node east-ocr crops all detected boxes
@@ -35,7 +35,7 @@ Clone github repository:
35
35
git clone https://github.com/argman/EAST
36
36
cd EAST
37
37
```
38
-
Download and unzip the file east_icdar2015_resnet_v1_50_rbox.zip to EAST folder with the github repository.
38
+
Download and unzip the file east_icdar2015_resnet_v1_50_rbox.zip as instructed in readme.md file to EAST folder with the github repository.
Copy file name to clipboardExpand all lines: docs/ensemble_scheduler.md
+13-16Lines changed: 13 additions & 16 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -25,7 +25,7 @@ This document presents ensemble models as an example of [DAG Scheduler](dag_sche
25
25
```
26
26
~$ virtualenv -p python3 .tf_env
27
27
~$ source .tf_env/bin/activate
28
-
~$ pip3 install tensorflow==1.15
28
+
~$ pip3 install tensorflow==2.3.1
29
29
```
30
30
3. Prepare argmax model with `(1, 1001)` input shapes to match output of googlenet and resnet output shapes. Generated model will sum inputs and calculate the index with the highest value. The model output will indicate the most likely predicted class from the ImageNet* dataset. <aname="point-3"></a>
31
31
```
@@ -39,56 +39,53 @@ This document presents ensemble models as an example of [DAG Scheduler](dag_sche
0 commit comments