Inference/Deploy
In this tab, you will see how to Inference and Deploy Trained Models in One Click from Deep Learning Studio.

Overview:

  • In this tab you will get four options :
1) Dataset Inference 2) Form Inference
3) Download 4) Deploy
1) Dataset Inference
  • When you click on the “Dataset Source” from “Dataset Inference” tab you will see 2 options i.e.
    • Validation :
      • You can use the Validation dataset for inference.
      • Select the Training Run.
      • And choose any output layer from your model or you can also choose default. You will see the option of all other layers by selecting the drop down button.
      • Click on Start inference.
      • Wait for a few secs, after that you will see the prediction output for the images.
    • Upload :
      • You can also upload your custom dataset file for inference.
      • Dataset upload instructions are given below the upload box in DLS.
      • Click on Start upload.
      • Wait for a few secs, after that you will see the prediction output for the images.
2) Form Inference
  • Choose Training run, and select or drop image here and click on Start inference.
3) Download
  • You can download the model in three ways :
    • Download Trained Model with Inference Code
      • Note : You can refer the "Download Model" link for detailed information on "How to download trained module with inference code and run inference"
    • Download Trained Model with Deployment Code
    • Download Trained Model Only
  • Select “Training Run” and download the model anyway by clicking on the options.
4) Deploy
  • You can refer to the Deployment link for detailed information.
Last modified 1yr ago
Copy link