From 047d35d0d8a2216a99a7664a3ee33e5cf6414e02 Mon Sep 17 00:00:00 2001 From: Jim Bennett Date: Sat, 26 Jun 2021 12:06:20 -0700 Subject: [PATCH] Assignment for object detector --- 5-retail/lessons/1-train-stock-detector/README.md | 6 +++--- 5-retail/lessons/1-train-stock-detector/assignment.md | 9 +++++++-- 2 files changed, 10 insertions(+), 5 deletions(-) diff --git a/5-retail/lessons/1-train-stock-detector/README.md b/5-retail/lessons/1-train-stock-detector/README.md index ada15c31..7a680a96 100644 --- a/5-retail/lessons/1-train-stock-detector/README.md +++ b/5-retail/lessons/1-train-stock-detector/README.md @@ -111,7 +111,7 @@ You can train an object detector using Custom Vision, in a similar way to how yo ![The settings for the custom vision project with the name set to fruit-quality-detector, no description, the resource set to fruit-quality-detector-training, the project type set to classification, the classification types set to multi class and the domains set to food](../../../images/custom-vision-create-object-detector-project.png) - > 💁 The products on shelves domain is specifically targeted for detecting stock on store shelves. + ✅ The products on shelves domain is specifically targeted for detecting stock on store shelves. Read more on the different domains in the [Select a domian documentation on Microsoft Docs](https://docs.microsoft.com/azure/cognitive-services/custom-vision-service/select-domain?WT.mc_id=academic-17441-jabenn#object-detection) ✅ Take some time to explore the Custom Vision UI for your object detector. @@ -119,7 +119,7 @@ You can train an object detector using Custom Vision, in a similar way to how yo To train your model you will need a set of images containing the objects you want to detect. -1. Gather images that contain the object to detect. You will need at least 15 images containing each object to detect from a variety of different angles and in different lighting conditions, but the more the better. You will also need a few images to test the model. If you are detecting more than one object, you will want some testing images that contain all the objects. +1. Gather images that contain the object to detect. You will need at least 15 images containing each object to detect from a variety of different angles and in different lighting conditions, but the more the better. This object detector uses the *Products on shelves* domain, so try to set up the objects as if they were on a store shelf. You will also need a few images to test the model. If you are detecting more than one object, you will want some testing images that contain all the objects. > 💁 Images with multiple different objects count towards the 15 image minimum for all the objects in the image. @@ -188,4 +188,4 @@ If you have any similar looking items, test it out by adding images of them to y ## Assignment -[](assignment.md) +[Compare domains](assignment.md) diff --git a/5-retail/lessons/1-train-stock-detector/assignment.md b/5-retail/lessons/1-train-stock-detector/assignment.md index da157d5c..13c342d0 100644 --- a/5-retail/lessons/1-train-stock-detector/assignment.md +++ b/5-retail/lessons/1-train-stock-detector/assignment.md @@ -1,9 +1,14 @@ -# +# Compare domains ## Instructions +When you created your object detector, you had a choice of multiple domains. Compare how well they work for your stock detector, and describe which gives better results. + +To change the domain, select the **Settings** button on the top menu, select a new domain, select the **Save changes** button, then retrain the model. Make sure you test with the new iteration of the model trained with the new domain. + ## Rubric | Criteria | Exemplary | Adequate | Needs Improvement | | -------- | --------- | -------- | ----------------- | -| | | | | +| Train the model with a different domain | Was able to change the domain and re-train the model | Was able to change the domain and re-train the model | Was unable to change the domain or re-train the model | +| Test the model and compare the results | Was able to test the model with different domains, compare results, and describe which is better | Was able to test the model with different domains, but was unable to compare the results and describe which is better | Was unable to test the model with different domains |