custom dataset train error
Opened this issue · 3 comments
Hi. I've tried demo image and found out that this SNIPER works very well.
I want to train this model with my own dataset.
Before trying, I wanted to check whether this training process works ok and I made a small datasets consists of 5 same images named, a1.jpg, a2.jpg, a3.jpg, a4.jpg, a5.jpg
And set annotation file as below which is in the same format to that of COCO.
However, error occurs when I run python main_train.py
####################
/home/wonjae/SNIPER/configs/faster/default_configs.py:171: YAMLLoadWarning: calling yaml.load() without Loader=... is deprecated, as the default Loader is unsafe. Please read https://msg.pyyaml.org/load for full details.
exp_config = edict(yaml.load(f))
loading annotations into memory...
Done (t=0.00s)
creating index...
index created!
num_images 5
COCO_train2014 gt roidb loaded from ./data/cache/COCO_train2014_gt_roidb.pkl
append flipped images to roidb
loading annotations into memory...
Done (t=0.00s)
creating index...
index created!
num_images 5
COCO_val2014 gt roidb loaded from ./data/cache/COCO_val2014_gt_roidb.pkl
append flipped images to roidb
filtered 2138 roidb entries: 246574 -> 244436
add bounding box regression targets
bbox target means:
[[0. 0. 0. 0.]
[0. 0. 0. 0.]]
[0. 0. 0. 0.]
bbox target stdevs:
[[0.1 0.1 0.2 0.2]
[0.1 0.1 0.2 0.2]]
[0.1 0.1 0.2 0.2]
Creating Iterator with 244436 Images
Total number of extracted chips: 955682
Done!
Traceback (most recent call last):
File "main_train.py", line 82, in
threads=config.TRAIN.NUM_THREAD, pad_rois_to=400)
File "lib/iterators/MNIteratorE2E.py", line 36, in init
super(MNIteratorE2E, self).init(roidb, config, batch_size, threads, nGPUs, pad_rois_to, False)
File "lib/iterators/MNIteratorBase.py", line 33, in init
self.get_batch()
File "lib/iterators/MNIteratorE2E.py", line 105, in get_batch
self.batch = self._get_batch()
File "lib/iterators/MNIteratorE2E.py", line 193, in _get_batch
processed_list = processed_list.get()
File "/home/wonjae/anacondas/envs/py27/lib/python2.7/multiprocessing/pool.py", line 572, in get
raise self._value
TypeError: 'NoneType' object has no attribute 'getitem'
######################
It seems that there are many others suffering from same problem.
I've tried the posted replies but it does not worksTT.
If anyone succeded in running his/her own datasets, can you please send the annotation file so that I can check out whether I am setting the annotation file right?
By the way, I've tried number of process/ number of thread to 8,4 but it did not work.
I will be looking forward for your reply.
Hi, have you solved this issue? I have encounter the same problem.
me too, the exactly same with you