Skip to content

Latest commit

 

History

History
94 lines (67 loc) · 4.1 KB

CONTRIBUTING.md

File metadata and controls

94 lines (67 loc) · 4.1 KB
Raw

Contributing

Setup

  • check out code
git clone -b ss-demo https://github.com/stevelr/wasmCloudArtefacts ml-demo
cd ml-demo
  • make sure you have the latest rust
rustup update
  • install cosmo-cli and wash-cli

Build actors and providers

From the top-level directory build with make. This should complete without errors.

Running locally-built components

It's best to start with a clean slate - no wasmcloud hosts running on your local machine

The script deploy/run-demo.sh contains commands to run everything. In the deploy folder, run run-demo.sh to see a list of available subcommands.

Start the local registry server, nats server, wasmcloud host, actors, and providers. If this is your first time running running this app, add --console to the end of the following command to open a new terminal window with the host logs. The logs may be useful for diagnosing any problems.

./run-demo.sh all
# or, to open a $TERMINAL window with host logs
./run-demo.sh all --console

After a successful startup, the inventory should look similar to the following:

wash get inventory

                                                                                         
                                          Host Inventory (NAZWBNEAMWM6B5EHTCRXKNI3LKYL3HFINT44CW7QRL3KVRH2UEJA4ALW)
                                                                                         
  hostcore.os                                                                            windows
  hostcore.osfamily                                                                      windows
  hostcore.arch                                                                          x86_64
  stargate                                                                               true
                                                                                         
  Actor ID                                                    Name                       Image Reference
  MDQTNL4LURMI5JKLFJNIVJQWQMGDJTXE7MWD7ASXP7PRFYAJQJBIYNG3    ML Inference UI            ghcr.io/liamrandall/image-ui:0.2.0
  MBZDEWD27BW6REESEOJ2EWT4GWVFCYQCYVS7D23Y53R5TUB65DIPV5UQ    ml_imagenetpreprocessor    ghcr.io/liamrandall/imagenetpreprocessor:0.1.0
  MAHR7434ZSCMRNY2T3W6AP3X2WZUJFSPHOF6HCB2EHSP3TTTZOVQJBZ3    ml_imagenetpostprocessor   ghcr.io/liamrandall/imagenetpostprocessor:0.1.0
  MBAS33Z3NX7K3OYDOIXORQHLUSAYMHDJB37B3HARJ27IJUWWQL46KUKR    ml_imagenetpreprocrgb8     ghcr.io/liamrandall/imagenetpreprocrgb8:0.1.0
  MA7ZIVAJMZSRVVKMFQQNSH4WQWKGE25MEUY7XRMGEHD3HCK5OAKWNJZH    inferenceapi               ghcr.io/connorsmith256/inferenceapi:0.2.0
                                                                                         
  Provider ID                                                 Name                       Link Name                Image Reference
  VBPGEQMDSZO35ERCCYMLMR3RQK7K5MMJPCSQNYOK37W4MQAEWUYDRSVZ    mlinference                default                  ghcr.io/connorsmith256/mlinference:0.2.1

If everything started correctly, try sending an image to be classified: (try any of the images in images/, or try one of your own!

curl -T images/cat.jpg http://localhost:8078/mobilenetv27/matches | jq

To stop the host and providers,

/run.sh wipe

Once the application is up and running, start to issue requests. Currently, the repository comprises the following pre-configured models:

  • identity of ONNX format
  • plus3 of Tensorflow format
  • mobilenet of ONNX format
  • squeezenet of ONNX format

Demo

  1. Select "priority:Accuracy", "model:Mobilenet v2.7"
  2. Then select a picture with the "Browse" button.
  3. Sample images are located in ml-demo/images
  4. Whenever the image is changed, it is sent through the recognition pipeline.
  5. One quirk of this page is that a new evaluation is only triggered on change of the image. Not any of the other fields, so if you want to use the same image with a different model, you need to change the model, then the image to something else, then change it back again.
  6. The resnet model almost always has higher accuracy than mobilenet.