Collapseruin commited on
Commit
20f7727
·
1 Parent(s): 3d3dedb

Change app.py file

Browse files
Files changed (1) hide show
  1. app.py +38 -21
app.py CHANGED
@@ -1,31 +1,48 @@
1
  import gradio as gr
2
  from PIL import Image
 
3
  import hopsworks
 
 
4
 
5
  project = hopsworks.login()
6
  fs = project.get_feature_store()
7
 
8
- dataset_api = project.get_dataset_api()
9
 
10
- dataset_api.download("Resources/images/latest_iris.png")
11
- dataset_api.download("Resources/images/actual_iris.png")
12
- dataset_api.download("Resources/images/df_recent.png")
13
- dataset_api.download("Resources/images/confusion_matrix.png")
 
14
 
15
- with gr.Blocks() as demo:
16
- with gr.Row():
17
- with gr.Column():
18
- gr.Label("Today's Predicted Image")
19
- input_img = gr.Image("latest_iris.png", elem_id="predicted-img")
20
- with gr.Column():
21
- gr.Label("Today's Actual Image")
22
- input_img = gr.Image("actual_iris.png", elem_id="actual-img")
23
- with gr.Row():
24
- with gr.Column():
25
- gr.Label("Recent Prediction History")
26
- input_img = gr.Image("df_recent.png", elem_id="recent-predictions")
27
- with gr.Column():
28
- gr.Label("Confusion Maxtrix with Historical Prediction Performance")
29
- input_img = gr.Image("confusion_matrix.png", elem_id="confusion-matrix")
 
 
 
 
 
 
 
 
 
 
 
 
 
 
30
 
31
- demo.launch()
 
1
  import gradio as gr
2
  from PIL import Image
3
+ import requests
4
  import hopsworks
5
+ import joblib
6
+ import pandas as pd
7
 
8
  project = hopsworks.login()
9
  fs = project.get_feature_store()
10
 
 
11
 
12
+ mr = project.get_model_registry()
13
+ model = mr.get_model("iris_model", version=1)
14
+ model_dir = model.download()
15
+ model = joblib.load(model_dir + "/iris_model.pkl")
16
+ print("Model downloaded")
17
 
18
+ def iris(sepal_length, sepal_width, petal_length, petal_width):
19
+ print("Calling function")
20
+ # df = pd.DataFrame([[sepal_length],[sepal_width],[petal_length],[petal_width]],
21
+ df = pd.DataFrame([[sepal_length,sepal_width,petal_length,petal_width]],
22
+ columns=['sepal_length','sepal_width','petal_length','petal_width'])
23
+ print("Predicting")
24
+ print(df)
25
+ # 'res' is a list of predictions returned as the label.
26
+ res = model.predict(df)
27
+ # We add '[0]' to the result of the transformed 'res', because 'res' is a list, and we only want
28
+ # the first element.
29
+ # print("Res: {0}").format(res)
30
+ print(res)
31
+ flower_url = "https://raw.githubusercontent.com/featurestoreorg/serverless-ml-course/main/src/01-module/assets/" + res[0] + ".png"
32
+ img = Image.open(requests.get(flower_url, stream=True).raw)
33
+ return img
34
+
35
+ demo = gr.Interface(
36
+ fn=iris,
37
+ title="Iris Flower Predictive Analytics",
38
+ description="Experiment with sepal/petal lengths/widths to predict which flower it is.",
39
+ allow_flagging="never",
40
+ inputs=[
41
+ gr.inputs.Number(default=2.0, label="sepal length (cm)"),
42
+ gr.inputs.Number(default=1.0, label="sepal width (cm)"),
43
+ gr.inputs.Number(default=2.0, label="petal length (cm)"),
44
+ gr.inputs.Number(default=1.0, label="petal width (cm)"),
45
+ ],
46
+ outputs=gr.Image(type="pil"))
47
 
48
+ demo.launch(debug=True)