Train.py could only train .png images, but not .jpg images

zw_sd · April 13, 2021, 2:10am

I am using jetson-inference/python/training/classification/train.py to train my model.

In the training images, there are .png and .jpg images.

I just found that the .jpg images was not trained.

Is there a quick way to make the .jpg be train with .png images?

Thank you.

AastaLLL · April 13, 2021, 5:07am

Hi,

jetson-inference uses torchvision for loading image data.

github.com

dusty-nv/pytorch-classification/blob/master/train.py#L164


      
              print("Use GPU: {} for training".format(args.gpu))
          
          
if args.distributed:
              if args.dist_url == "env://" and args.rank == -1:
                  args.rank = int(os.environ["RANK"])
              if args.multiprocessing_distributed:
                  # For multiprocessing distributed training, rank needs to be the
                  # global rank among all the processes
                  args.rank = args.rank * ngpus_per_node + gpu
              dist.init_process_group(backend=args.dist_backend, init_method=args.dist_url,
                                      world_size=args.world_size, rank=args.rank)
          
          
# data loading code
          traindir = os.path.join(args.data, 'train')
          valdir = os.path.join(args.data, 'val')
          normalize = transforms.Normalize(mean=[0.485, 0.456, 0.406],
                                           std=[0.229, 0.224, 0.225])
          
          
train_dataset = datasets.ImageFolder(
              traindir,
              transforms.Compose([

Based on the document below, torchvision should support both .jpg and .png file extension:
https://github.com/pytorch/vision/blob/master/torchvision/datasets/folder.py#L234

IMG_EXTENSIONS = ('.jpg', '.jpeg', '.png', '.ppm', '.bmp', '.pgm', '.tif', '.tiff', '.webp')

Could you validate if the data is stored in the recommended placement first:
https://pytorch.org/vision/stable/datasets.html

root/dog/xxx.png
root/dog/xxy.png
root/dog/[...]/xxz.png

root/cat/123.png
root/cat/nsdf3.png
root/cat/[...]/asd932_.png

Thanks.

zw_sd · April 13, 2021, 8:08am

thank you @AastaLLL for your quickly response.
for now I converted .jpg files to .png.
I just found that maybe just because I made a mistake when detecting the images. :P
Let me check further. :)

AastaLLL · April 14, 2021, 8:18am

Sure. Thanks for the update !

Topic		Replies	Views
Training on jetson-inference getting error FileNotFoundError: Found no valid file for the classes Drowsy. Supported extensions are: .jpg, .jpeg, Jetson Nano	4	1287	December 8, 2021
Running into some issues with training a new dataset with jetson-inference Frameworks pytorch	0	458	October 28, 2020
Jetson inference "collecting your own detection datasets" Deep Learning (Training & Inference) jetson-inference	1	625	November 3, 2020
Jetson Inference Train Issue TensorRT	1	505	December 14, 2020
Train_ssd.py error - Training Object Detection Models Jetson Nano ai-training	10	1415	October 6, 2022
How to Retrain an image Classifier for New Categories Jetson TX2	7	458	October 18, 2021
Help with image recognition Jetson Nano ai-training	2	476	October 18, 2021
I cannot installing PyTorch for python2.7 Frameworks pytorch	0	463	May 29, 2020
Googlenet network's place Jetson Nano jetson-inference	25	1256	October 18, 2021
Re-training ResNet-18 Model Jetson TX2 jetson-inference	6	628	October 18, 2021

Train.py could only train .png images, but not .jpg images

Related topics