Error is keep occuring --- AssertionError: Empty dataset

[ysssssha]

I tried to train for stpls3d and s3dis dataset with the same procedure of data preparation, however, the error of empty dataset is shown as the result.
This is the code and results. actually i should input the '/' to the end of data path during '3) Preprocess data' in data preparation.
How can i solve this problems?
Thanks

stpls3d

s3dis

for the stpls3d

(isbnet) ha@ha-Z790-AORUS-ELITE:~/Downloads/ISBNet-master$ python3 tools/train.py configs/stpls3d/isbnet_stpls3d.yaml --trainall --exp_name default
2024-03-26 15:25:24,731 - INFO - Train all !!!!!!!!!!!!!!!!
2024-03-26 15:25:24,731 - INFO - Config:
model:
channels: 16
num_blocks: 7
semantic_classes: 15
instance_classes: 14
sem2ins_classes: []
semantic_only: False
semantic_weight: [1.0, 1.0, 44.0, 21.9, 1.8, 25.1, 31.5, 21.8, 24.0, 54.4, 114.4, 81.2, 43.6, 9.7, 22.4]
with_coords: False
ignore_label: -100
voxel_scale: 3
use_spp_pool: False
filter_bg_thresh: 0.1
iterative_sampling: False
mask_dim_out: 32
instance_head_cfg:
num_dyco_layer: 2
dec_dim: 64
n_sample_pa1: 2048
n_queries: 256
radius_scale: 10
radius: 0.4
neighbor: 16
test_cfg:
x4_split: False
logit_thresh: 0.0
score_thresh: 0.2
npoint_thresh: 10
type_nms: 'matrix'
topk: 100

fixed_modules: ['input_conv', 'unet', 'output_layer', 'semantic_linear', 'offset_linear', 'offset_vertices_linear', 'box_conf_linear']

data:
train:
type: 'stpls3d'
data_root: 'dataset/stpls3d/train/'
prefix: 'train'
suffix: '_inst_nostuff.pth'
training: True
repeat: 3
voxel_cfg:
scale: 3
spatial_shape: [128, 512]
max_npoint: 250000
min_npoint: 5000
test:
type: 'stpls3d'
data_root: 'dataset/stpls3d/val_250m/'
prefix: 'val'
suffix: '_inst_nostuff.pth'
training: False
voxel_cfg:
scale: 3
spatial_shape: [128, 512]
max_npoint: 250000
min_npoint: 5000

dataloader:
train:
batch_size: 32
num_workers: 16
test:
batch_size: 1
num_workers: 1

optimizer:
type: 'AdamW'
lr: 0.001
weight_decay: 0.0001

save_cfg:
semantic: False
offset: False
instance: True
offset_vertices: False
nmc_clusters: False
object_conditions: False

fp16: True
epochs: 120
step_epoch: 50
save_freq: 4
pretrain: 'pretrains/stpls3d/pretrain_stpls3d_val.pth'
work_dir: ''

2024-03-26 15:25:24,731 - INFO - Distributed: False
2024-03-26 15:25:24,731 - INFO - Mix precision training: True
2024-03-26 15:25:24,732 - INFO - Save at: ./work_dirs/stpls3d/isbnet_stpls3d/default
2024-03-26 15:25:25,333 - INFO - Total params: 7713104
2024-03-26 15:25:25,333 - INFO - Trainable params: 7713104
Traceback (most recent call last):
File "tools/train.py", line 303, in
main()
File "tools/train.py", line 260, in main
train_set = build_dataset(cfg.data.train, logger)
File "/home/ha/Downloads/ISBNet-master/isbnet/data/init.py", line 25, in build_dataset
return STPLS3DDataset(**_data_cfg)
File "/home/ha/Downloads/ISBNet-master/isbnet/data/custom.py", line 26, in init
self.filenames = self.get_filenames()
File "/home/ha/Downloads/ISBNet-master/isbnet/data/custom.py", line 38, in get_filenames
assert len(filenames) > 0, "Empty dataset."
AssertionError: Empty dataset.

for the s3dis

(isbnet) ha@ha-Z790-AORUS-ELITE:~/Downloads/ISBNet-master$ python3 tools/train.py configs/s3dis/isbnet_s3dis_area5.yaml --trainall --exp_name default
2024-03-26 15:33:07,276 - INFO - Train all !!!!!!!!!!!!!!!!
2024-03-26 15:33:07,276 - INFO - Config:
model:
channels: 32
num_blocks: 7
semantic_classes: 13
instance_classes: 11
sem2ins_classes: [0,1]
semantic_only: False
semantic_weight: False
with_coords: True
ignore_label: -100
voxel_scale: 50
use_spp_pool: True
filter_bg_thresh: 0.4
iterative_sampling: True
instance_head_cfg:
num_dyco_layer: 3
dec_dim: 128
n_sample_pa1: 2048
n_queries: 256
radius_scale: 1
radius: 0.4
neighbor: 32
test_cfg:
x4_split: True
logit_thresh: 0.0
score_thresh: 0.5
npoint_thresh: 400
type_nms: 'standard'
topk: -1
nms_threshold: 0.2

fixed_modules: ['input_conv', 'unet', 'output_layer', 'semantic_linear', 'offset_linear', 'offset_vertices_linear', 'box_conf_linear']

data:
train:
type: 's3dis'
data_root: 'dataset/s3dis/'
prefix: ['Area_1', 'Area_2', 'Area_3', 'Area_4', 'Area_6']
suffix: '_inst_nostuff.pth'
repeat: 20
training: True
voxel_cfg:
scale: 50
spatial_shape: [128, 512]
max_npoint: 250000
min_npoint: 5000
test:
type: 's3dis'
data_root: 'dataset/s3dis/'
prefix: 'Area_5'
suffix: '_inst_nostuff.pth'
training: False
voxel_cfg:
scale: 50
spatial_shape: [128, 512]
max_npoint: 250000
min_npoint: 5000

dataloader:
train:
batch_size: 12
num_workers: 12
test:
batch_size: 1
num_workers: 1

optimizer:
type: 'AdamW'
lr: 0.001
weight_decay: 0.0001

save_cfg:
semantic: False
offset: False
instance: True
offset_vertices: False
nmc_clusters: False
object_conditions: False

fp16: False
epochs: 120
step_epoch: 50
save_freq: 4
pretrain: 'pretrains/s3dis/pretrain_s3dis_area5.pth'
work_dir: ''
2024-03-26 15:33:07,276 - INFO - Distributed: False
2024-03-26 15:33:07,276 - INFO - Mix precision training: False
2024-03-26 15:33:07,277 - INFO - Save at: ./work_dirs/s3dis/isbnet_s3dis_area5/default
2024-03-26 15:33:07,974 - INFO - Total params: 30720427
2024-03-26 15:33:07,974 - INFO - Trainable params: 30720427
[]
Traceback (most recent call last):
File "tools/train.py", line 303, in
main()
File "tools/train.py", line 260, in main
train_set = build_dataset(cfg.data.train, logger)
File "/home/ha/Downloads/ISBNet-master/isbnet/data/init.py", line 19, in build_dataset
return S3DISDataset(**_data_cfg)
File "/home/ha/Downloads/ISBNet-master/isbnet/data/custom.py", line 26, in init
self.filenames = self.get_filenames()
File "/home/ha/Downloads/ISBNet-master/isbnet/data/s3dis.py", line 36, in get_filenames
assert len(filenames) > 0, f"Empty {p}"
AssertionError: Empty Area_1

[Smith-Augmentus]

It appears that the code is not handling files with the .pth extension correctly. For example, the code currently doesn't check for files with a .pth extension, but the input files contain .pth extensions. This discrepancy could lead to issues when loading or saving the data.

Suggested Solution:

Please consider updating the code to properly account for files with the .pth extension, or alternatively, ensure that the files are renamed with the appropriate extension before processing.