search

mahyarnajibi / SNIPER

SNIPER / AutoFocus is an efficient multi-scale object detection training / inference algorithm
Other
2.69k stars 450 forks source link

Unable to run demo_3k.py #75

Open RajathSwaroop opened 5 years ago

RajathSwaroop commented 5 years ago

Traceback (most recent call last): File "demo_3k.py", line 21, in from bbox.bbox_transform import bbox_pred, clip_boxes File "lib/bbox/bbox_transform.py", line 2, in from bbox import bbox_overlaps_cython, ignore_overlaps_cython ImportError: cannot import name bbox_overlaps_cython

This is the error I get after installing mxnet and other pre-requisites and running the demo_3k python executable.

Also I would like to know if there is a build process for running the R-FCN 3000 on a GPU machine

RajathSwaroop commented 5 years ago

Solved the issue by running "sudo python setup.py build_ext --inplace" inside the /lib/bbox and /lib/dataset/pycocotools directories

however stuck at

File "demo.py", line 52, in main sym = sym_inst.get_symbol_rcnn(config, is_train=False) File "/home/rajath/work/SNIPER-cvpr3k/symbols/faster/resnet_mx_101_e2e_3k_demo.py", line 212, in get_symbol_rcnn scales=(tuple)(cfg.network.ANCHOR_SCALES) ) ValueError: need more than 1 value to unpack

lishuchao012 commented 5 years ago

Met the same problem . Did you solve it ?

Traceback (most recent call last): File "demo.py", line 244, in main() File "demo.py", line 52, in main sym = sym_inst.get_symbol_rcnn(config, is_train=False) File "/home/master/datas/deep-learning/sniper/cvpr3k/symbols/faster/resnet_mx_101_e2e_3k_demo.py", line 202, in get_symbolrcnn rois, = mx.sym.MultiProposal(cls_prob=rpn_cls_prob_reshape, bbox_pred=rpn_bbox_pred, im_info=im_info, AttributeError: 'module' object has no attribute 'MultiProposal'