AttributeError: 'list' object has no attribute 'reshape'. for big frames - numpy

Question

this is my code. and i don't know what is the problem. the my goal is read more number frames and then convert them to array and finally split them to train and test data.
def read_frames(PATH, num1, num2,seq_len,count,num):
directory = os.listdir(PATH)
for i in range(num1,num2):
source_folder = PATH + '/' +'{}'.format(directory[i])
print("appending the images is started for:", directory[i])
for item in sorted(os.listdir(source_folder), key = len):
img = Image.open(os.path.join(source_folder, item))
frames.append(item)
x.append(np.asarray(img))
count+=1
if count == seq_len:
break
this is a function that read the frames from the path and convert them to array. seq_len is the number of frames.i read the frames in 3 parts. 2 parts using the function that mentioned and other part is:
seq_len = 1624
frames = []
x = []
y = []
w=[]
count = 0
num = 0
PATH = ""
directory = os.listdir(PATH)
print("#.................reading and appending the frames for PART1 are started...........#","\n")
for i in range(0, len(directory)):
source_folder = "PATH/{}".format(directory[i])
print(num)
# print(source_folder)
print("appending the images is started for:", directory[i])
for item in sorted(os.listdir(source_folder), key = len):
# print(item)
# print(count)
img = Image.open(os.path.join(source_folder, item))
frames.append(item)
x.append(np.asarray(img))
count+=1
if count == seq_len:
break
the total number of frames that reading from 3 parts are 333060.
when running this cell:
w = np.array(x)
print('w_shape: ', np.array(w, dtype='uint8').shape)
i expected that the output be (333060,224,224,3) but it is not. and i face this error.
ValueError: setting an array element with a sequence.
when the number of frames are not more this error doesn't happening and when the number of frames are more i face with that error. i using this frames for my pre-trained cnn + lstm network.
please help me to solve that.