GeorgeSeif / Semantic-Segmentation-Suite

Semantic Segmentation Suite in TensorFlow. Implement, train, and test new Semantic Segmentation models easily!
2.51k stars 880 forks source link

Pretrained weights not available #201

Closed jesarte closed 5 years ago

jesarte commented 5 years ago

Describe the problem

Some of the pretrained weights for the frontends are not available. I have tried all frontends: ResNet50, ResNet152, MobileNetV2 and InceptionV4, and all fail at the same point.

Source code / logs

Preparing the model ... --2019-03-22 09:44:10-- https://s3.amazonaws.com/pretrained-weights/mobilenet_v2.ckpt.meta Resolving s3.amazonaws.com (s3.amazonaws.com)... 54.231.114.44 Connecting to s3.amazonaws.com (s3.amazonaws.com)|54.231.114.44|:443... connected. HTTP request sent, awaiting response... 404 Not Found 2019-03-22 09:44:11 ERROR 404: Not Found.

Traceback (most recent call last): File "utils/get_pretrained_checkpoints.py", line 21, in subprocess.check_output(['wget','https://s3.amazonaws.com/pretrained-weights/mobilenet_v2.ckpt.meta', "-P", "models"]) File "/usr/lib/python3.6/subprocess.py", line 336, in check_output kwargs).stdout File "/usr/lib/python3.6/subprocess.py", line 418, in run output=stdout, stderr=stderr) subprocess.CalledProcessError: Command '['wget', 'https://s3.amazonaws.com/pretrained-weights/mobilenet_v2.ckpt.meta', '-P', 'models']' returned non-zero exit status 8. Traceback (most recent call last): File "./train.py", line 100, in network, init_fn = model_builder.build_model(model_name=args.model, frontend=args.frontend, net_input=net_input, num_classes=num_classes, crop_width=args.crop_width, crop_height=args.crop_height, is_training=True) File "/content/gdrive/My Drive/Sync/Semantic-Segmentation-Suite/builders/model_builder.py", line 51, in build_model download_checkpoints("MobileNetV2") File "/content/gdrive/My Drive/Sync/Semantic-Segmentation-Suite/builders/model_builder.py", line 28, in download_checkpoints subprocess.check_output(["python", "utils/get_pretrained_checkpoints.py", "--model=" + model_name]) File "/usr/lib/python3.6/subprocess.py", line 336, in check_output kwargs).stdout File "/usr/lib/python3.6/subprocess.py", line 418, in run output=stdout, stderr=stderr) subprocess.CalledProcessError: Command '['python', 'utils/get_pretrained_checkpoints.py', '--model=MobileNetV2']' returned non-zero exit status 1.

iambackit commented 5 years ago

Describe the problem

Some of the pretrained weights for the frontends are not available. I have tried all frontends: ResNet50, ResNet152, MobileNetV2 and InceptionV4, and all fail at the same point.

Source code / logs

Preparing the model ... --2019-03-22 09:44:10-- https://s3.amazonaws.com/pretrained-weights/mobilenet_v2.ckpt.meta Resolving s3.amazonaws.com (s3.amazonaws.com)... 54.231.114.44 Connecting to s3.amazonaws.com (s3.amazonaws.com)|54.231.114.44|:443... connected. HTTP request sent, awaiting response... 404 Not Found 2019-03-22 09:44:11 ERROR 404: Not Found.

Traceback (most recent call last): File "utils/get_pretrained_checkpoints.py", line 21, in subprocess.check_output(['wget','https://s3.amazonaws.com/pretrained-weights/mobilenet_v2.ckpt.meta', "-P", "models"]) File "/usr/lib/python3.6/subprocess.py", line 336, in check_output kwargs).stdout File "/usr/lib/python3.6/subprocess.py", line 418, in run output=stdout, stderr=stderr) subprocess.CalledProcessError: Command '['wget', 'https://s3.amazonaws.com/pretrained-weights/mobilenet_v2.ckpt.meta', '-P', 'models']' returned non-zero exit status 8. Traceback (most recent call last): File "./train.py", line 100, in network, init_fn = model_builder.build_model(model_name=args.model, frontend=args.frontend, net_input=net_input, num_classes=num_classes, crop_width=args.crop_width, crop_height=args.crop_height, is_training=True) File "/content/gdrive/My Drive/Sync/Semantic-Segmentation-Suite/builders/model_builder.py", line 51, in build_model download_checkpoints("MobileNetV2") File "/content/gdrive/My Drive/Sync/Semantic-Segmentation-Suite/builders/model_builder.py", line 28, in download_checkpoints subprocess.check_output(["python", "utils/get_pretrained_checkpoints.py", "--model=" + model_name]) File "/usr/lib/python3.6/subprocess.py", line 336, in check_output kwargs).stdout File "/usr/lib/python3.6/subprocess.py", line 418, in run output=stdout, stderr=stderr) subprocess.CalledProcessError: Command '['python', 'utils/get_pretrained_checkpoints.py', '--model=MobileNetV2']' returned non-zero exit status 1.

I have the same issue.

jsolves commented 5 years ago

The Commit on Mar 7, 2019 has updated urls (that may work)...

jesarte commented 5 years ago

Those seem to work! Thank you very much!

yucheng-china commented 4 years ago

I have tried to use the new url, but this issue still happend, could you mind tell me how to slove it ?

iambackit commented 4 years ago

I have tried to use the new url, but this issue still happend, could you mind tell me how to slove it ?

Sadly I couldn't solve that problem, so I found another project.