CODEWITHSUNDEEP
tensorflowdeep-learningkeras
Aprameyo Roy
Aprameyo Roy

Reputation: 1899

Keras Top 5 predictions

How can I retrieve the top 5 predictions from model.predict() in Keras? It only gives only 1 prediction. Is there any way to do so? I don't want it as an evaluation metric. I just need the top 5 predictions.

Upvotes: 7

Views: 5297

Answers (3)

Joy
Joy

Reputation: 41

You can try with the below code.

n = 5
y_preds = self.model.predict(x)
y_preds = np.argsort(y_preds, axis=1)[:,-n:]

Upvotes: 1

JtotheR
JtotheR

Reputation: 503

if labels are the indices of the correct class:

top1 = 0.0
top5 = 0.0    
class_probs = model.predict(x)
for i, l in enumerate(labels):
    class_prob = class_probs[i]
    top_values = (-class_prob).argsort()[:5]
    if top_values[0] == l:
        top1 += 1.0
    if np.isin(np.array([l]), top_values):
        top5 += 1.0

print("top1 acc", top1/len(labels))
print("top1 acc", top5/len(labels))

Upvotes: 0

thefifthjack005
thefifthjack005

Reputation: 638

if you are trying to get top prediction from a image classification problem, you will receive a one hot code prediction.

class_prob = [0.98,0.50,0.60,0.90,0.87,0.79,0.87]
top_values_index = sorted(range(len(class_prob)), key=lambda i: class_prob[i])[-the_top_values_you_want_to_extract:]

you now have the index for all the five top values.you can now just loop through the index and get the class name.

to extract just the top_values_without_index

top_values= [class_prob[i] for i in np.argsort(class_prob)[-5:]]

Upvotes: 4

Related Questions