Update img

Author: curiousily

manuscript/03.image-classification-with-transfer-learning.md

@@ -600,7 +600,7 @@ Ok, but how good our model will be when confronted with a real-world image? Let'
 show_image('stop-sign.jpg')
 ```
 
-![png](./stop-sign.jpg)
+![jpg](./stop-sign.jpg)
 
 For this, we'll have a look at the confidence for each class. Let's get this from our model:
 
@@ -657,7 +657,7 @@ The last challenge for our model is a traffic sign that it hasn't seen before:
 show_image('unknown-sign.jpg')
 ```
 
-![png](./unknown-sign.jpg)
+![jpg](./unknown-sign.jpg)
 
 Let's get the predictions:
 
@@ -782,7 +782,7 @@ Again, our model is learning very quickly. Let's have a look at the sample image
 show_image('unknown-sign.jpg')
 ```
 
-![png](./unknown-sign.jpg)
+![jpg](./unknown-sign.jpg)
 
 ```py
 pred = predict_proba(enchanced_model, 'unknown-sign.jpg')