Upload app.py

app.py

@@ -1,46 +1,27 @@
 import gradio as gr
-
-
-
-
+import tensorflow as tf
 import numpy as np
-from transformers import TFLiteModel
-
-# Load the TFLite model from Hugging Face
-model = TFLiteModel.from_pretrained('https://huggingface.co/spaces/JefferyJapheth/sega/blob/main/model.tflite')
-
-# Define preprocessing function for .npy files
-def preprocess_frame(file_path):
-    # Load the .npy file
-    frame_data = np.load(file_path)
 
-    # Replace this with any preprocessing needed for your model
-    # For example, normalization, reshaping, etc.
-    processed_frame = frame_data  # Replace this with your actual preprocessing code
-    return processed_frame
+# Load your sign language prediction model
+model = tf.keras.models.load_model(r"C:\Users\Jeffery.st\Desktop\st.Jeffery00\mega\model.h5")
 
-# Provide the fixed file path for the .npy input
-npy_file_path = "https://huggingface.co/spaces/JefferyJapheth/sega/blob/main/NON_EMPTY_FRAME_IDXS.npy"
-
-# Define the predict_sign_language function for Gradio interface
-def predict_sign_language(webcam_frame):
-    # Preprocess the .npy file
-    processed_frame = preprocess_frame(npy_file_path)
-
-    # Process the webcam frame (convert Gradio's image format to numpy array if needed)
-    webcam_frame_np = webcam_frame
-
-    # Combine the frames, or perform any required processing
-
-    # Run inference using the TFLite model
-    prediction = model(processed_frame)
+# Define the prediction function
+def predict_sign_language(image):
+    # Run inference using the model
+    prediction = model.predict(np.expand_dims(image, axis=0))
 
     # ... (post-processing if required)
 
     return prediction
 
-# Create the Gradio interface with a webcam input
-webcam_input = gr.inputs.Image(shape=(224, 224))  # Set the shape according to your model's requirements
+# Define the webcam input component
+webcam_input = gr.inputs.Image(source="webcam")
+
+# Define the output component
+label_output = gr.outputs.Label()
+
+# Create the Gradio interface
+iface = gr.Interface(fn=predict_sign_language, inputs=webcam_input, outputs=label_output)
 
-iface = gr.Interface(fn=predict_sign_language, inputs=webcam_input, outputs="text")
-iface.launch(share=True)  # Set `share=True` to make it accessible outside the localhost
+# Launch the interface
+iface.launch()