train icdar2003 meet some problem

[gittigxuy]

When I train the icdar 2003 dataset,I have generated a caffemodel,but during the test period,I meet this problem:

Traceback (most recent call last): File "./tools/test_net.py", line 86, in imdb = get_imdb(args.imdb_name) File "/home/xuy/code/RRPN/tools/../lib/datasets/factory.py", line 37, in get_imdb raise KeyError('Unknown dataset: {}'.format(name)) KeyError: 'Unknown dataset: MSRA_TEST'

according to your faster_rcnn_end2end.sh,I could know that:

case $DATASET in pascal_voc) TRAIN_IMDB="voc_2007_trainval" TEST_IMDB="voc_2007_test" PT_DIR="pascal_voc" ITERS=70000 ;; coco)

This is a very long and slow training schedule

# You can probably use fewer iterations and reduce the
# time to the LR drop (set in the solver to 350,000 iterations).
TRAIN_IMDB="coco_2014_train"
TEST_IMDB="coco_2014_minival"
PT_DIR="coco"
ITERS=490000
;;

rrpn)

This is a very long and slow training schedule

# You can probably use fewer iterations and reduce the
# time to the LR drop (set in the solver to 350,000 iterations).
TRAIN_IMDB="MSRA_TRAIN"
TEST_IMDB="MSRA_TEST"
PT_DIR="rrpn"
ITERS=490000
;;

rrpn_vehicle)

This is a very long and slow training schedule

# You can probably use fewer iterations and reduce the
# time to the LR drop (set in the solver to 350,000 iterations).
TRAIN_IMDB="MSRA_TRAIN"
TEST_IMDB="MSRA_TEST"
PT_DIR="rrpn_vehicle"
ITERS=490000
;;

*) echo "No dataset given" exit ;; esac

so,what should I do?

[ghost]

Hello @gittigxuy. I met the same problem also. As much as I understood the evaluation part has not released yet. Nevertheless there are some other issues opened with the same problem. @mjq11302010044 do you plan to release that part? I think following the same procedure will be better for acceptable comparision.

Thanks.

[mjq11302010044]

@gittigxuy @gyaliniz We haven't release a general evaluation API yet, but demo code can be use for reference in tools/joint_test.py.