aws / amazon-sagemaker-examples

Example 📓 Jupyter notebooks that demonstrate how to build, train, and deploy machine learning models using 🧠 Amazon SageMaker.
https://sagemaker-examples.readthedocs.io
Apache License 2.0
9.97k stars 6.73k forks source link

managed_spot_training_object_detection.ipynd fatal error in loading data #1612

Open duane-edgington opened 3 years ago

duane-edgington commented 3 years ago

Hi. I am trying to walk through the 10-minute tutorial on Optimizing and Scaling Machine Learning Training. When I try to run my copy of the managed-spot_training_object_detection.ipynd notebook, following the tutorial steps exactly, I get a fatal error at Download and Prepare Data. --2020-10-12 23:50:32-- http://host.robots.ox.ac.uk/pascal/VOC/voc2012/VOCtrainval_11-May-2012.tar Resolving host.robots.ox.ac.uk (host.robots.ox.ac.uk)... 129.67.94.152 Connecting to host.robots.ox.ac.uk (host.robots.ox.ac.uk)|129.67.94.152|:80... failed: No route to host. --2020-10-12 23:50:35-- http://host.robots.ox.ac.uk/pascal/VOC/voc2007/VOCtrainval_06-Nov-2007.tar Resolving host.robots.ox.ac.uk (host.robots.ox.ac.uk)... 129.67.94.152 Connecting to host.robots.ox.ac.uk (host.robots.ox.ac.uk)|129.67.94.152|:80... failed: No route to host. --2020-10-12 23:50:38-- http://host.robots.ox.ac.uk/pascal/VOC/voc2007/VOCtest_06-Nov-2007.tar Resolving host.robots.ox.ac.uk (host.robots.ox.ac.uk)... 129.67.94.152 Connecting to host.robots.ox.ac.uk (host.robots.ox.ac.uk)|129.67.94.152|:80... failed: No route to host. tar: /tmp/VOCtrainval_11-May-2012.tar: Cannot open: No such file or directory tar: Error is not recoverable: exiting now tar: /tmp/VOCtrainval_06-Nov-2007.tar: Cannot open: No such file or directory tar: Error is not recoverable: exiting now tar: /tmp/VOCtest_06-Nov-2007.tar: Cannot open: No such file or directory tar: Error is not recoverable: exiting now Traceback (most recent call last): File "tools/prepare_dataset.py", line 95, in db = load_pascal(args.set, args.year, args.root_path, args.shuffle) File "tools/prepare_dataset.py", line 63, in load_pascal imdbs.append(PascalVoc(s, y, devkit_path, shuffle, is_train=True)) File "/home/ec2-user/SageMaker/managed_spot_training_object_detection_2020-10-12/tools/pascal_voc.py", line 59, in init self.image_set_index = self._load_image_set_index(shuffle) File "/home/ec2-user/SageMaker/managed_spot_training_object_detection_2020-10-12/tools/pascal_voc.py", line 91, in _load_image_set_index assert os.path.exists(image_set_index_file), 'Path does not exist: {}'.format(image_set_index_file) AssertionError: Path does not exist: /home/ec2-user/SageMaker/managed_spot_training_object_detection_2020-10-12/tools/../VOCdevkit/VOC2007/ImageSets/Main/trainval.txt Traceback (most recent call last): File "tools/prepare_dataset.py", line 95, in db = load_pascal(args.set, args.year, args.root_path, args.shuffle) File "tools/prepare_dataset.py", line 63, in load_pascal imdbs.append(PascalVoc(s, y, devkit_path, shuffle, is_train=True)) File "/home/ec2-user/SageMaker/managed_spot_training_object_detection_2020-10-12/tools/pascal_voc.py", line 59, in init self.image_set_index = self._load_image_set_index(shuffle) File "/home/ec2-user/SageMaker/managed_spot_training_object_detection_2020-10-12/tools/pascal_voc.py", line 91, in _load_image_set_index assert os.path.exists(image_set_index_file), 'Path does not exist: {}'.format(image_set_index_file) AssertionError: Path does not exist: /home/ec2-user/SageMaker/managed_spot_training_object_detection_2020-10-12/tools/../VOCdevkit/VOC2007/ImageSets/Main/test.txt

I will go look at the data repository website to see if there is an update to the files, but in the meantime please check the notebook and correct the data download so us users can make progress.

Thank yoiu.

duane-edgington commented 3 years ago

I checked host.robots.ox.ac.uk and that site is not responding. Has been down for several days. I found an alternative mirror for the three data sets, and that worked. http://pjreddie.com/media/files/VOCtrainval_11-May-2012.tar and the other two tar files as well. exactly the same names, and same size. I also found a mirror at Baidu but the file names were different (encapsulated in a zip file I believe). Anyway, it would be great if you could update the notebooks with working repositories, even better if they were within AWS.