woltium
commented
2 years ago like:
when i try python main.py
it turns out:
/home/wws/anaconda3/envs/transloc/lib/python3.7/site-packages/MinkowskiEngine-0.5.4-py3.7-linux-x86_64.egg/MinkowskiEngine/init.py:42: UserWarning: The environment variable OMP_NUM_THREADS not set. MinkowskiEngine will automatically set OMP_NUM_THREADS=16. If you want to set OMP_NUM_THREADS manually, please export it on the command line before running a python script. e.g. export OMP_NUM_THREADS=12; python your_program.py. It is recommended to set it below 24.
"It is recommended to set it below 24.",
2022-05-04 20:56:42,541 - INFO - Config :
optimizer_type = 'Adam'
optimizer_cfg = dict(lr=0.0002, weight_decay=0, betas=(0.9, 0.999))
scheduler_type = 'MultiStepLR'
scheduler_cfg = dict(gamma=0.1, milestones=(80, 120, 160))
end_epoch = 200
loss_type = 'BatchHardTripletMarginLoss'
loss_cfg = dict(margin=0.2, normalize_embeddings=False)
train_cfg = dict(
save_per_epoch=2,
val_per_epoch=2,
batch_sampler_type='ExpansionBatchSampler',
batch_sampler_cfg=dict(
max_batch_size=256,
batch_size_expansion_rate=1.4,
batch_expansion_threshold=0.7,
batch_size=32,
shuffle=True,
drop_last=True),
num_workers=16)
eval_cfg = dict(
batch_sampler_cfg=dict(batch_size=16, drop_last=False),
num_workers=16,
normalize_embeddings=False)
model_type = 'TransLoc3D'
model_cfg = dict(
backbone_cfg=dict(
up_conv_cfgs=[
dict(in_channels=1, out_channels=64, kernel_size=5, stride=1),
dict(in_channels=64, out_channels=64, kernel_size=2, stride=2)
],
transformer_cfg=dict(
num_attn_layers=6,
global_channels=64,
local_channels=0,
num_centers=[256, 128, 128, 64, 64, 64],
num_heads=2),
in_channels=1,
out_channels=512),
pool_cfg=dict(
type='NetVlad',
in_channels=512,
out_channels=256,
cluster_size=64,
gating=True,
add_bn=True),
quantization_size=0.01)
dataset_type = 'Oxford'
data_root_dir = '/home/wws/PycharmProjects/pythonProject/TransLoc3D-main/datasets/preprocess/'
dataset_cfg = dict(
data_root_dir=
'/home/wws/PycharmProjects/pythonProject/TransLoc3D-main/datasets/preprocess/',
transform_cfg=dict(
train=[
dict(
type='Jitter3D',
objs=['pcd'],
params=dict(sigma=0.001, clip=0.002)),
dict(
type='Drop3D',
objs=['pcd'],
params=dict(method='random', min_dr=0.0, max_dr=0.1)),
dict(
type='Translation3D',
objs=['pcd'],
params=dict(max_delta=0.01)),
dict(
type='Drop3D',
objs=['pcd'],
params=dict(method='cuboid', p=0.4, bbox_pc_name='pcd')),
dict(type='ToTensor', objs=['pcd'], params=dict())
],
database=[dict(type='ToTensor', objs=['pcd'], params=dict())],
queries=[dict(type='ToTensor', objs=['pcd'], params=dict())]),
batch_transform_cfg=dict(
train=[
dict(
type='Rotation3D',
objs=['pcd'],
params=dict(
batch=True, max_theta1=5, max_theta2=0, axis=[0, 0, 1])),
dict(
type='Mirror3D',
objs=['pcd'],
params=dict(
batch=True, method='xyz_plane', p=[0.25, 0.25, 0.0]))
],
database=[],
queries=[]),
train_catalog_file_path=
'/home/wws/PycharmProjects/pythonProject/TransLoc3D-main/datasets/preprocess/training_queries_baseline.pickle',
cached_train_catalog_file_path=
'/home/wws/PycharmProjects/pythonProject/TransLoc3D-main/datasets/preprocess/training_queries_baseline_cached.pickle',
database_file_path=
'/home/wws/PycharmProjects/pythonProject/TransLoc3D-main/datasets/preprocess/oxford_evaluation_database.pickle',
queries_file_path=
'/home/wws/PycharmProjects/pythonProject/TransLoc3D-main/datasets/preprocess/oxford_evaluation_query.pickle'
)
task_type = 'default_me'
work_dir = '/home/wws/PycharmProjects/pythonProject/TransLoc3D-main/workspace/transloc3d_baseline_cfg/2022-05-04_20-56-41'
resume_from = None
resume_items = [
'model', 'epoch', 'metrics', 'optim', 'sched', 'sampler', 'all',
'no_metrics'
]
mode = 'train'
debug = False
seed = None
deterministic = False
2022-05-04 20:56:42,695 - INFO - Building default_me task ...
1
2
3
2022-05-04 20:56:42,809 - INFO - Parameters in default_me task : 10972111
cuda:0
2022-05-04 20:56:44,713 - INFO - [Epoch 1/ 200] Start training ...
0%| | 0/679 [00:02<?, ?it/s]
Traceback (most recent call last):
File "main.py", line 101, in
slothfulxtx
Thank you for your code, which gave me a lot of help, but I found that MinkowskiEngine always made mistakes in the process of trying to run the code. What version of MinkowskiEngine do you use?