Thanks so much for providing this excellent code base. My team at the US Geological Survey is using it to recognize character on historical topographic maps. When training, the ICDAR datasets throw an index out of range error in segmentation_mask. If we don't use those datasets (set ratio to zero in the config files), training proceeds normally. We downloaded those datasets from the place you indicate, and have them in the right spot with a train_list.txt file as directed. Please advise us if possible.
Traceback (most recent call last):
File "tools/train_net.py", line 174, in
main()
File "tools/train_net.py", line 167, in main
model = train(cfg, args.local_rank, args.distributed)
File "tools/train_net.py", line 66, in train
do_train(
File "/data/sams_projects/SpotElevationsOCR/MaskTextSpotter/maskrcnn_benchmark/engine/trainer.py", line 66, in do_train
loss_dict = model(images, targets)
File "/data/sams_projects/conda_envs/mts/lib/python3.8/site-packages/torch/nn/modules/module.py", line 532, in call
result = self.forward(*input, kwargs)
File "/data/sams_projects/SpotElevationsOCR/MaskTextSpotter/maskrcnn_benchmark/modeling/detector/generalized_rcnn.py", line 52, in forward
x, result, detector_losses = self.roi_heads(features, proposals, targets)
File "/data/sams_projects/conda_envs/mts/lib/python3.8/site-packages/torch/nn/modules/module.py", line 532, in call
result = self.forward(*input, *kwargs)
File "/data/sams_projects/SpotElevationsOCR/MaskTextSpotter/maskrcnn_benchmark/modeling/roi_heads/roi_heads.py", line 36, in forward
x, detections, loss_mask = self.mask(mask_features, detections, targets)
File "/data/sams_projects/conda_envs/mts/lib/python3.8/site-packages/torch/nn/modules/module.py", line 532, in call
result = self.forward(input, kwargs)
File "/data/sams_projects/SpotElevationsOCR/MaskTextSpotter/maskrcnn_benchmark/modeling/roi_heads/mask_head/mask_head.py", line 185, in forward
mask_targets, char_mask_targets, char_mask_weights, decoder_targets, word_targets = self.prepare_targets(proposals, targets)
File "/data/sams_projects/SpotElevationsOCR/MaskTextSpotter/maskrcnn_benchmark/modeling/roi_heads/mask_head/mask_head.py", line 124, in prepare_targets
matched_targets = self.match_targets_to_proposals(
File "/data/sams_projects/SpotElevationsOCR/MaskTextSpotter/maskrcnn_benchmark/modeling/roi_heads/mask_head/mask_head.py", line 113, in match_targets_to_proposals
matched_targets = target[matched_idxs.clamp(min=0)]
File "/data/sams_projects/SpotElevationsOCR/MaskTextSpotter/maskrcnn_benchmark/structures/bounding_box.py", line 246, in getitem
bbox.add_field(k, v[item])
File "/data/sams_projects/SpotElevationsOCR/MaskTextSpotter/maskrcnn_benchmark/structures/segmentation_mask.py", line 501, in getitem
selected_chars_boxes.append(self.chars_boxes[i])
IndexError: list index out of range
Traceback (most recent call last):
File "/data/sams_projects/conda_envs/mts/lib/python3.8/runpy.py", line 193, in _run_module_as_main
return _run_code(code, main_globals, None,
File "/data/sams_projects/conda_envs/mts/lib/python3.8/runpy.py", line 86, in _run_code
exec(code, run_globals)
File "/data/sams_projects/conda_envs/mts/lib/python3.8/site-packages/torch/distributed/launch.py", line 263, in
main()
File "/data/sams_projects/conda_envs/mts/lib/python3.8/site-packages/torch/distributed/launch.py", line 258, in main
raise subprocess.CalledProcessError(returncode=process.returncode,
subprocess.CalledProcessError: Command '['/data/sams_projects/conda_envs/mts/bin/python3', '-u', 'tools/train_net.py', '--local_rank=0', '--config-file', 'configs/finetune.yaml']' returned non-zero exit
Hi @MhLiao,
Thanks so much for providing this excellent code base. My team at the US Geological Survey is using it to recognize character on historical topographic maps. When training, the ICDAR datasets throw an index out of range error in segmentation_mask. If we don't use those datasets (set ratio to zero in the config files), training proceeds normally. We downloaded those datasets from the place you indicate, and have them in the right spot with a train_list.txt file as directed. Please advise us if possible.
Thanks so much,
Sam