I am using your code to run on my custom dataset. However, I am getting nan on Many_shpt_accuracy_top1 on val set. I tried changing the learning rate too, but it doesn't seem to work.
{'criterions': {'PerformanceLoss': {'def_file': 'loss/SoftmaxLoss.py',
'loss_params': {},
'optim_params': None,
'weight': 1.0}},
'memory': {'centroids': False, 'init_centroids': False},
'networks': {'classifier': {'def_file': 'models/DotProductClassifier.py',
'optim_params': {'lr': 1,
'momentum': 0.9,
'weight_decay': 0},
'params': {'dataset': 'ImageNet_LT',
'in_dim': 512,
'num_classes': 201,
'stage1_weights': False}},
'feat_model': {'def_file': 'models/ResNet10Feature.py',
'fix': False,
'optim_params': {'lr': 1,
'momentum': 0.1,
'weight_decay': 0},
'params': {'caffe': False,
'dataset': 'ImageNet_LT',
'dropout': None,
'stage1_weights': False,
'use_fc': False,
'use_modulatedatt': False}}},
'training_opt': {'batch_size': 128,
'dataset': 'ImageNet_LT',
'display_step': 10,
'feature_dim': 512,
'log_dir': 'logs/ImageNet_LT/stage1',
'num_classes': 201,
'num_epochs': 500,
'num_workers': 4,
'open_threshold': 0.01,
'sampler': None,
'scheduler_params': {'gamma': 0.1, 'step_size': 1000}}}
Loading data from /pylon5/pscstaff/rajanie/MLStamps/long-tail/OpenLongTailRecognition-OLTR/OLTRDataset/OLTRDataset_1/train.txt
Use data transformation: Compose(
RandomResizedCrop(size=(224, 224), scale=(0.08, 1.0), ratio=(0.75, 1.3333), interpolation=PIL.Image.BILINEAR)
RandomHorizontalFlip(p=0.5)
ColorJitter(brightness=0.4, contrast=0.4, saturation=0.4, hue=0)
ToTensor()
Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
)
No sampler.
Shuffle is True.
Loading data from /pylon5/pscstaff/rajanie/MLStamps/long-tail/OpenLongTailRecognition-OLTR/OLTRDataset/OLTRDataset_1/val.txt
Use data transformation: Compose(
Resize(size=256, interpolation=PIL.Image.BILINEAR)
CenterCrop(size=(224, 224))
ToTensor()
Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
)
No sampler.
Shuffle is True.
Using 1 GPUs.
Loading Scratch ResNet 10 Feature Model.
No Pretrained Weights For Feature Model.
Loading Dot Product Classifier.
Random initialized classifier weights.
Using steps for training.
Initializing model optimizer.
Loading Softmax Loss.
Hi,
I am using your code to run on my custom dataset. However, I am getting nan on Many_shpt_accuracy_top1 on val set. I tried changing the learning rate too, but it doesn't seem to work.
Can you give me any pointers to debug this? I didn't make any changes to the code.