Merge pull request #26 from datarootsio/bugfix

Bugfix

Author: baturayo

README.md

@@ -110,7 +110,8 @@ Follow the steps below, to run the entire pipeline.
 ```
 git clone https://github.com/datarootsio/face-mask-detection.git
 cd face-mask-detection
-tar xvfz data.tar.gz
+mkdir data
+tar -xvf data.tar.gz -C ./data
 pip install -r requirements.txt
 ```
 After completing these steps, you can run and play with the model in `scripts/predict.ipynb` notebook.
@@ -124,32 +125,9 @@ If you want to recreate the `data` folder and retrain `masked or not masked` cla
   5. Run `predict.ipynb` to run entire pipeline and see an example output of face mask detection model.
 
 ## Getting Started for Calling Deployed Face Mask Detection Model
-The model has been deployed in the [dploy.ai](dploy.ai) platform. By making a REST call, you can provide your image and get the prediction response.
-
-The input should have the following format:
-
-```
-{
- image: <Base64 Image String>,
- type: <Image Type e.g. jpg, jpeg, png>
-}
-```
-The response will have the following format:
-```
-{
-    'detected_face_coordinates': <the bounding box coordinates of the detected faces e.g.
-                                 [[x1, y1, x2, y2], [x1, y1, x2, y2]]>,
-    'detected_mask_scores': <the prediction score of the detected faces between 0 and 1 e.g.
-                            ["0.8", "0.99", "0.001"]>,
-    'detected_face_labels': <human readible classification results of the detected faces e.g.
-                            [masked, masked, not masked]>,
-    'annotated_image': <annotated base64 image that visualizes the prediction results with  
-                       bounding boxes and text labels>,
-    'image_type': <Image Type e.g. jpg, jpeg, png>
-}
-```
-
+The model is ready to be deployed using [dploy.ai](dploy.ai) platform. `dploy.ai` is a brand new deployment platform where you can dploy your ML models super easily. Please follow [`this guide`](https://docs.dploy.ai/docs) to deploy an example model.
 
+You can also easily deploy this repo locally using the open sourced [`dploy-kickstart`](https://docs.dploy.ai/docs/python#4-testing-dploy-locally).
 
 ## Contact
 Ping us:

data.tar.gz

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:536f2fab0a94bfaef794a4b2b75faa490df5f313b628f1e9401270283af99c08
-size 103918001
+oid sha256:f8056eabf9d7c7b36917053ca57e21985c5ee1b4afe8d5fe89dc6f7c7355555a
+size 106224675

dploy.yaml

@@ -1,29 +1,8 @@
 model_id: d36795f9-0376-4a4a-bc05-0fb964ee229b
-description: |
-  Face masks are crucial in minimizing the propagation of COVID-19, and are highly recommended or even obligatory in many situations.
-
-  This deployed model gives access to dataroots' [open-sourced](https://github.com/datarootsio/face-mask-detection) face mask detection project.
-
-  You can interactively test the model below. Or you can use the model via through our API with the following steps:
-
-  1. Subscribe to this model
-  2. Copy your user-id and api-key tokens
-  3. Perform the following request
-
-  ```sh
-  curl --request POST \
-    --url https://${modelversion_id}.users.dploy.ai/ \
-    --header 'content-type: application/json' \
-    --header 'x-api-key: <insert your api key>' \
-    --header 'x-api-user: <insert your api id>' \
-    --data '{"image": "<insert your base64string>", "type": "png"}'
-  ```
-  For an example of how to use this within your application, check out our [examples](https://github.com/dploy-ai/dploy-examples)
-spec_version: "0.1"
+spec_version: v2
 spec:
   environment:
     base: python:3.7
     entrypoint: scripts/predict.ipynb
     requirements: ["requirements.txt"]
   replicas: 1
-  max_replicas: 5

requirements.txt

@@ -1,4 +1,4 @@
-tensorflow==2.1.2
+tensorflow==2.3
 numpy==1.18.2
 opencv-python==4.2.0.34
 PyYAML==5.3.1
@@ -14,4 +14,4 @@ googledrivedownloader==0.4
 face-detection==0.1.4
 tqdm==4.44.1
 pandas==1.0.3
-face_detection==0.1.4
+face-recognition==1.3.0