CODEWITHSUNDEEP
deep-learningpytorchconfusion-matrix
Itsmrshow
Itsmrshow

Reputation: 28

I am having trouble calculating the accuracy, recall, precision and f1-score for my model

I have got my confusion matrix working correctly, just having some trouble producing the scores. A little help would go a long way. I am currently getting the error. "Tensor object is not callable". 

def get_confused(model_ft):
    nb_classes = 120
    from sklearn.metrics import precision_recall_fscore_support as score
    confusion_matrix = torch.zeros(nb_classes, nb_classes)
    with torch.no_grad():
        for i, (inputs, classes) in enumerate(dataloaders['val']):
            inputs = inputs.to(device)
            classes = classes.to(device)
            outputs = model_ft(inputs)
            _, preds = torch.max(outputs, 1)
            for t, p in zip(classes.view(-1), preds.view(-1)):
                    confusion_matrix[t.long(), p.long()] += 1

            cm = confusion_matrix(classes, preds)
            recall = np.diag(cm) / np.sum(cm, axis = 1)
            precision = np.diag(cm) / np.sum(cm, axis = 0)
    print(confusion_matrix)
    print(confusion_matrix.diag()/confusion_matrix.sum(1))

Upvotes: 0

Views: 1749

Answers (2)

Srikanth Reddy
Srikanth Reddy

Reputation: 121

You can try the below code, 

def F_score(logit, label, threshold=0.5, beta=2):
    prob = torch.sigmoid(logit)
    prob = prob > threshold
    label = label > threshold
    TP = (prob & label).sum().float()
    TN = ((~prob) & (~label)).sum().float()
    FP = (prob & (~label)).sum().float()
    FN = ((~prob) & label).sum().float()
    accuracy = (TP+TN)/(TP+TN+FP+FN)
    precision = torch.mean(TP / (TP + FP + 1e-12))
    recall = torch.mean(TP / (TP + FN + 1e-12))
    F2 = (1 + beta**2) * precision * recall / (beta**2 * precision + recall + 1e-12)
    return accuracy, precision, recall, F2.mean(0)

call the funciton as,

accuracy, precision, recall, F1_score = F_score(output.squeeze(), labels.float())

Reference:- https://www.kaggle.com/c/human-protein-atlas-image-classification/discussion/73246

Upvotes: 2

Umang Gupta
Umang Gupta

Reputation: 16450

The problem is with this line. 

cm = confusion_matrix(classes, preds)

confusion_matrix is a tensor and you can't call it like a function. Hence Tensor is not callable. I am also, not sure why you need this line. Instead, I think you would want to write cm= confusion_matrix.cpu().data.numpy() to make it a numpy array I think. From your code, it seems cm is np.array.

Upvotes: 2

Related Questions