Create app.py

app.py

@@ -0,0 +1,54 @@
+import tensorflow as tf
+import keras
+import numpy as np
+import cv2
+import random
+
+classes = ['paper', 'rock', 'scissors']
+
+model = tf.keras.models.load_model('rps.h5')
+
+def classify_image(inp):
+  img = cv2.resize(inp,(224,224),interpolation=cv2.INTER_AREA)
+  img = np.reshape(img,(1,224,224,3))
+  pred = model1.predict(img).flatten()
+  confidences = {classes[i]: float(pred[i]) for i in range(3)}
+  global det
+  det = classes[pred.argmax(axis=-1)]
+  print(det)
+  return confidences
+
+def random_char(n):
+  print("HELLO")
+  n = random.randint(0,2)
+  global sel
+  sel = classes[n]
+  print(sel)
+  return classes[n].upper()
+
+def result(t):
+  print("HIOAL")
+  print(sel, det)
+  if (sel == 'rock' and det == 'paper') or (sel == 'paper' and det == 'scissors') or (sel == "scissors" and det == "rock"):
+    return "YOU WON"
+  elif (sel == 'paper' and det == 'rock') or (sel == 'scissors' and det == 'paper') or (sel == "rock" and det == "scissors"):
+    return "YOU LOST"
+  else:
+    return "IT'S A TIE"
+
+import gradio as gr
+
+webcam = gr.inputs.Image(shape=(224, 224), source="webcam")
+classify = gr.Interface(fn=classify_image, 
+             inputs=webcam,
+             outputs=gr.Label(num_top_classes=3))
+
+computer = gr.Interface(fn=random_char,
+                        inputs=None,
+                        outputs=gr.TextArea(max_lines=1,label='The computer selected:'))
+
+final = gr.Interface(fn=result,
+                        inputs=None,
+                        outputs=gr.TextArea(max_lines=1,label='Result'))
+
+final = gr.Parallel(classify, computer, final).launch(debug=True)