Upload 2 files

app.py

@@ -2,49 +2,39 @@ import gradio as gr
 import tensorflow as tf
 import gdown
 import numpy as np
-from transformers import TFAutoModelForImageClassification
-import torch
+from keras.models import load_model
 
-path = [['0229_01.jpg'], ['0385_01.jpg'], ['0067_01.jpg']]
+path  = [['0229_01.jpg'], ['0385_01.jpg'], ['0067_01.jpg']]
 
 URL = 'https://drive.google.com/file/d/1TpQmvq2R8nHy9CQaVb0jtMMuRQA0Cw0L/view?usp=drive_link'
 output_path = 'Labels_train_abd.txt'
-gdown.download(URL, output_path, quiet=False, fuzzy=True)
+gdown.download(URL, output_path, quiet=False,fuzzy=True)
 
-with open(output_path, 'r') as file:
+with open(output_path,'r') as file:
     LABELS = [x.strip() for x in file.readlines()]
 
 num_classes = 4000
 IMG_SIZE = 128
 
 def _normalize_img(img):
-    img = tf.cast(img, tf.float32) / 255.  # All images will be rescaled by 1./255
-    img = tf.image.resize(img, (IMG_SIZE, IMG_SIZE), method='bilinear')
-    return img
+    img = tf.cast(img, tf.float32)/255.   # All images will be rescaled by 1./255
+    img = tf.image.resize(img,  (IMG_SIZE, IMG_SIZE), method= 'bilinear')
+    return (img)
 
-# Load your Hugging Face model
-model = TFAutoModelForImageClassification.from_pretrained("final_model.h5")
+model = load_model("final_model.h5")
 
 def predict_fn(img):
-    img = img.convert('RGB')
-    img_data = _normalize_img(img)
-    x = np.array(img_data)
-    x = np.expand_dims(x, axis=0)
-    
-    # Convert the input to a torch tensor
-    inputs = torch.tensor(x)
-
-    # Perform inference using the Hugging Face model
-    with torch.no_grad():
-        outputs = model(inputs)
-    
-    # Get the predicted labels and their scores
-    predicted_labels = torch.argsort(outputs.logits[0], descending=True)[:3]
-    top3_scores = outputs.logits[0][predicted_labels]
-    
-    # Map the labels to their corresponding names
-    top3_labels = [LABELS[i] for i in predicted_labels]
-    
-    return {label: str(score) for label, score in zip(top3_labels, top3_scores)}
-
-gr.Interface(predict_fn, gr.inputs.Image(type='pil'), outputs="label", examples=path).launch()
+     img = img.convert('RGB')
+     img_data = _normalize_img(img)
+     x = np.array(img_data)
+     x = np.expand_dims(x, axis=0)
+     temp = model.predict(x)
+     
+     idx = np.argsort(np.squeeze(temp))[::-1]
+     top3_value = np.asarray([temp[0][i] for i in idx[0:3]])
+     top3_idx = idx[0:3]
+     
+     return {LABELS[i]:str(v) for i,v in zip(top3_idx,top3_value)}
+
+
+gr.Interface(predict_fn, gr.inputs.Image(type='pil'), outputs='label', examples=path,).launch()

requirements.txt

@@ -9,7 +9,6 @@ scipy>=1.4.1
 gradio>=3.36.1
 tensorflow==2.12.0
 tensorflow-datasets==4.9.2
-transformers==4.12.0
 
 
 # Plotting ------------------------------------