SVM classifies all samples as one class

Hey guys,

I’m working in a personal project which is a simple classifier 0 or 1.
I have traied two different algorithms.
First of all Logistic Regression which obtains the following results:
precision recall f1-score support

     0.0       0.78      0.70      0.73       417
     1.0       0.67      0.76      0.71       342

     accuracy                           0.72       759
   macro avg       0.72      0.73      0.72       759
weighted avg       0.73      0.72      0.72       759

Confusion Matrix
[[290 127]
 [ 83 259]]

1

Next I splited again the dataset and I tried with the SVM algorithm.
the problem is that when I do the predictions with the x_test splited dataset all values are classified as 0.0.
Here I the confusion matrix and the classification report obtained for this algorithm.
precision recall f1-score support

     0.0       0.55      1.00      0.71       417
     1.0       0.00      0.00      0.00       342

     accuracy                           0.55       759
   macro avg       0.27      0.50      0.35       759
weighted avg       0.30      0.55      0.39       759

Confusion Matrix
[[417   0]
 [342   0]]

The follwing warning I supose it occurs because is dividing the 0.00 with the 0.55.

What may be happening? I suppose it’s a comon error.