问题描述
使用tensorboard显示图像出现问题
--------------------------------------------------------------------------- AssertionError Traceback (most recent call last) <ipython-input-96-71fff8c3a347> in <module> 2 grid = torchvision.utils.make_grid(img) 3 ----> 4 logger.add_images('images', grid, 0) D:\Anaconda\lib\site-packages\torch\utils\tensorboard\writer.py in add_images(self, tag, img_tensor, global_step, walltime, dataformats) 599 img_tensor = workspace.FetchBlob(img_tensor) 600 self._get_file_writer().add_summary( --> 601 image(tag, img_tensor, dataformats=dataformats), global_step, walltime) 602 603 def add_image_with_boxes(self, tag, img_tensor, box_tensor, global_step=None, D:\Anaconda\lib\site-packages\torch\utils\tensorboard\summary.py in image(tag, tensor, rescale, dataformats) 394 """ 395 tensor = make_np(tensor) --> 396 tensor = convert_to_HWC(tensor, dataformats) 397 # Do not assume that user passes in values in [0, 255], use data type to detect 398 scale_factor = _calc_scale_factor(tensor) D:\Anaconda\lib\site-packages\torch\utils\tensorboard\_utils.py in convert_to_HWC(tensor, input_format) 99 input_format: {}".format(input_format) 100 assert(len(tensor.shape) == len(input_format)), "size of input tensor and input format are different. \ --> 101 tensor shape: {}, input_format: {}".format(tensor.shape, input_format) 102 input_format = input_format.upper() 103 AssertionError: size of input tensor and input format are different. tensor shape: (3, 138, 274), input_format: NCHW
解决方案：
logger.add_image('images', grid, 0)
