mobileSAM: RuntimeError: Non-empty 4D data tensor expected but got a tensor with sizes [1, 0, 576, 1024]

Question

when I try to run the following code on Ubuntu with python = 3.11.9, pytorch = 2.3.1+cu121

from ultralytics import SAM
import cv2
import time
import numpy as np
import torch

model = SAM("mobile_sam.pt")  # Load model
capture = cv2.VideoCapture('images/0-1.MOV')  

t = []
while capture.isOpened():
    ret, frame = capture.read()
    if not ret:
        break
    frame = cv2.resize(frame, (640,480)) 
    print('start interence... , ', type(frame))
    t1 = time.time()
    results = model.predict(frame) # Inference
    print(f'done inference... , {time.time()-t1} ', type(results))
    t.append(time.time()-t1)
    
capture.release()

I got an error like this:

Traceback (most recent call last):
  File "/home/ubuntu/SLAM/SLAM/test.py", line 28, in <module>
    results = model.predict(frame) # Inference
              ^^^^^^^^^^^^^^^^^^^^
RuntimeError: Non-empty 4D data tensor expected but got a tensor with sizes [1, 0, 576, 1024]

However, I run the same code successfully on my mac. And I want to find how to solve this probelm.