The docker image is hosted on sagecontinuum.
Build the image:
docker build -t sagecontinuum/plugin-smokedetect:ai-gateway-demo .
Run the container:
docker run sagecontinuum/plugin-smokedetect:ai-gateway-demo --siteID 0 --cameraType 0
Example output of the plugin:
Get image from HPWREN Camera
Image url: http://hpwren.ucsd.edu/cameras/L/bm-n-mobo-c.jpg
Description: Big Black Mountain North Color Original
Perform an inference based on trainned model
Fire, 71.29%
Publishmic pkg start --name smoke-detection
After this point mic will take you to the docker container that it created to encapsulate the model. MIC will ask for a framework/language. For this model, pick generic since its using python3.6 and this model uses python3.6.
Select a option <enum 'Framework'> (generic, python3.7, python3.8, conda4.7.12): generic
After this point, mic takes you to the docker enviroment to setup the encapsulation:
pip3 install -r requirements.txt
mic pkg trace python3 main.py --cameraType 0 --siteID 0
mic pkg parameters -f mic/mic.yaml -n siteID -v 0 -d 'Input parameter used to provide the hpwren camera site'
mic pkg parameters -f mic/mic.yaml -n cameraType -v 0 -d 'Input parameter used to provide the hpwren camera type'
mic pkg inputs
mic pkg outputs
mic pkg wrapper
mic pkg run
If all the commands were successful, mic will prompt you to exit the docker container by typing exit.
Now we are going to upload the model to MIC (assumes the right credentials are already set).
mic pkg upload
mic will prompt you with information about the model.
Check with dame (assumes wifire profile is set in credentials):
dame run a365022a-8ccb-47e5-89aa-3560c0cf8f2e -p wifire
dame will prompt you with the following questions and with the answers:
To run this model configuration,a hpwren_py file (.py file) is required.
Please enter a url: hpwren.py
To run this model configuration,a inference_py file (.py file) is required.
Please enter a url: inference.py
To run this model configuration,a model_tflite file (.tflite file) is required.
Please enter a url: model.tflite
To run this model configuration,a __pycache___zip file (.zip file) is required.
Please enter a url: https://s3.mint.mosorio.dev/components/mint_component_20210609-133724.zip
Next you can edit the default parameters or not:
Do you want to edit the parameters? [y/N]: N
Finally you can run the docker container and get your results:
Do you want to proceed and submit it for execution? [Y/n]: Y
First need to prepare your binder repository Next, expose software inputs and outputs. This is already done for this notebook. [Convert the repository to a software component] (https://mic-cli.readthedocs.io/en/latest/notebooks/convert_repository/):
mic notebook read https://github.com/iperezx/edge-plugin-smokedetect
The command should generate the following main.cwl file:
arguments:
- --
baseCommand: /app/cwl/bin/main
class: CommandLineTool
cwlVersion: v1.1
hints:
DockerRequirement:
dockerImageId: r2d-2fvar-2ffolders-2fj6-2ff847737s17s4hwjc1g8n46600000gn-2ft-2frepo2cwl-5fsa7v7oym-2frepo1623273231
inputs:
cameraType:
inputBinding:
prefix: --cameraType
type: int
modelPath:
inputBinding:
prefix: --modelPath
type: File
siteID:
inputBinding:
prefix: --siteID
type: int
outputs:
imagePath:
outputBinding:
glob: ./hpwren-image-used-for-inference.jpeg
type: File
resultsPath:
outputBinding:
glob: ./model-inference-results.json
type: File
requirements:
NetworkAccess:
networkAccess: true
Test the configuration file with a values.yaml file:
cameraType: 0
siteID: 0
modelPath:
class: File
path: /Users/iperezx/Documents/edge-plugin-smokedetect/model.tflite
cwltool main.cwl values.yaml
Expected output should be:
... Starting smoke detection inferencing Get image from HPWREN Camera Image url: http://hpwren.ucsd.edu/cameras/L/tje-1-mobo-c.jpg Description: Unknown direction Color Original Perform an inference based on trainned model Fire, 62.04% INFO [job main.cwl] Max memory used: 108MiB INFO [job main.cwl] completed success { "imagePath": { "location": "file:///Users/iperezx/Documents/edge-plugin-smokedetect/hpwren-image-used-for-inference.jpeg", "basename": "hpwren-image-used-for-inference.jpeg", "class": "File", "checksum": "sha1$e98f05c7871a47e3543c91b75bcb1a5efcefbdb3", "size": 3300, "path": "/Users/iperezx/Documents/edge-plugin-smokedetect/hpwren-image-used-for-inference.jpeg" }, "resultsPath": { "location": "file:///Users/iperezx/Documents/edge-plugin-smokedetect/model-inference-results.json", "basename": "model-inference-results.json", "class": "File", "checksum": "sha1$221b0f02811f658726bc116c446b6e1e9df559ff", "size": 283, "path": "/Users/iperezx/Documents/edge-plugin-smokedetect/model-inference-results.json" } }
Finally the upload part:
Upload Docker Image:
mic notebook upload-image main.cwl
Upload Model Component
mic notebook upload-component main.cwl values.yaml
Now test with `dame`:
dame run 9b2d70e9-e6f9-4ac4-9dff-62b8e1991080 -p wifire
Prompts you for the path of the `model.tflite`:
To run this model configuration,a modelPath file (.unknown file) is required. Please enter a url: model.tflite
`dame` will generate multiple .yaml files to run now with `cwltool`:
cwltool /Users/iperezx/Documents/edge-plugin-smokedetect/spec.yaml /Users/iperezx/Documents/edge-plugin-smokedetect/values.yml
If successfull, this is the expected output:
... Starting smoke detection inferencing Get image from HPWREN Camera Image url: http://hpwren.ucsd.edu/cameras/L/tje-1-mobo-c.jpg Description: Unknown direction Color Original Perform an inference based on trainned model Fire, 67.10% INFO [job spec.yaml] Max memory used: 108MiB INFO [job spec.yaml] completed success { "imagePath": { "location": "file:///Users/iperezx/Documents/edge-plugin-smokedetect/hpwren-image-used-for-inference.jpeg", "basename": "hpwren-image-used-for-inference.jpeg", "class": "File", "checksum": "sha1$bea7edd55cf9b2c9a81ceaebc7c5213b5fb5fd2c", "size": 3266, "path": "/Users/iperezx/Documents/edge-plugin-smokedetect/hpwren-image-used-for-inference.jpeg" }, "resultsPath": { "location": "file:///Users/iperezx/Documents/edge-plugin-smokedetect/model-inference-results.json", "basename": "model-inference-results.json", "class": "File", "checksum": "sha1$b6f4b4cf5a16f4ffa6f52dc0b62ee48503285144", "size": 283, "path": "/Users/iperezx/Documents/edge-plugin-smokedetect/model-inference-results.json" } } INFO Final process status is success