Download and Deploy your DLS trained model on your machine
Once your are done with the training you will be able to download the deployment code. To download , follow the steps below:
- 1.Go to Inference/Deploy in your project
- 2.Click on "Download trained model with deployment code"
It will download a zip files which contains the trained model and deployment server packages. Now follow the next section to use the downloaded file.
- 1.Extract the downloaded zip file.
- 2.Launch command prompt(Windows) or terminal(Linux/Ubuntu)
- 3.change the directory to the extract path
cd <EXTACTED_TRAINED_PACKAGE> - 4.Create conda enviromentNote : In the older version yaml file name may be different , please use it accordinglya) for windows cpu version:
conda env create -f env_windows_cpu.yaml -n deployment_envb) for windows gpu version:conda env create -f env_windows_gpu.yaml -n deployment_envc) for linux cpu version:conda env create -f env_linux_cpu.yaml -n deployment_envd) for linux gpu version:conda env create -f env_linux_gpu.yaml -n deployment_envNote : If you don't find any environment yaml please download it from the one of this link- - 5.Activate the environment
conda activate deployment_env - 6.Set the deployment directoryIf you are using terminal or bash:
export DEPLOY_DIR=./If you using windows cmd:set DEPLOY_DIR=.\ - 7.Start the deployment serverolympus up --no-debug --port 11111 --host 127.0.0.1 &olympus deploy --name MNISTHandwrittenDigitsClassifierUsingRNN_Run0_deploy_model --version 1 --framework keras ./model/
You have sucessfully deployed to model model deployed at http://127.0.0.1:11111/models/MNISTHandwrittenDigitsClassifierUsingRNN_Run0_deploy_model/v1/app.html
- If you are using any cloud machine/server/instance for the deployment, You need to connect to the ssh and follow the steps 3-6.