[Python] how to convert a PIL image to a pytorch tensor for py-feat

i need to convert a pil image into a pytorch tensor to use it in Detector.detect() (py-feat) but i got this error in line where i call detect():

RuntimeError: Given groups=1, weight of size [64, 3, 3, 3], expected input[1, 1, 112, 112] to have 3 channels, but got 1 channels instead
​

however, if i check how tensor looks like using check_tensor() after converting, it looks just fine

here's the code:

def check_tensor(self, frame_tensor):
frame_numpy = frame_tensor.squeeze(0).permute(1, 2, 0).byte().numpy()
frame_numpy_bgr = cv2.cvtColor(frame_numpy, cv2.COLOR_RGB2BGR)
cv2.imshow("Frame", frame_numpy_bgr)
cv2.waitKey(0)
cv2.destroyAllWindows()

def analyze_image(self, image_path):

img = Image.open(image_path)
tensor_image = F.pil_to_tensor(img)
orig_prediction = self.detector.detect(inputs=tensor_image, data_type="tensor")


how can i fix this?

Continue reading...