Hello, is there a tool for generating point annotation json file in the current folder?

Environment ( I have tried pytorch 1.9.0 and pytorch 1.8.0 and their corresponding cudnn/ mmcv-full =1.14.0, both will cause memory increasing when training.)

sys.platform: linux
Python: 3.8.17 (default, Jul 5 2023, 21:04:15) [GCC 11.2.0]
CUDA available: True
GPU 0: NVIDIA TITAN RTX
CUDA_HOME: /usr/local/cuda-10.2
NVCC: Cuda compilation tools, release 10.2, V10.2.89
GCC: gcc (Ubuntu 7.5.0-3ubuntu1~18.04) 7.5.0
PyTorch: 1.10.0
PyTorch compiling details: PyTorch built with:

• GCC 7.3
• C++ Version: 201402
• Intel(R) oneAPI Math Kernel Library Version 2023.1-Product Build 20230303 for Intel(R) 64 architecture applications
• Intel(R) MKL-DNN v2.2.3 (Git Hash 7336ca9f055cf1bfa13efb658fe15dc9b41f0740)
• OpenMP 201511 (a.k.a. OpenMP 4.5)
• LAPACK is enabled (usually provided by MKL)
• NNPACK is enabled
• CPU capability usage: AVX512
• CUDA Runtime 11.3
• NVCC architecture flags: -gencode;arch=compute_37,code=sm_37;-gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_61,code=sm_61;-gencode;arch=compute_70,code=sm_70;-gencode;arch=compute_75,code=sm_75;-gencode;arch=compute_80,code=sm_80;-gencode;arch=compute_86,code=sm_86;-gencode;arch=compute_37,code=compute_37
• CuDNN 8.2
• Magma 2.5.2
• Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, CUDA_VERSION=11.3, CUDNN_VERSION=8.2.0, CXX_COMPILER=/opt/rh/devtoolset-7/root/usr/bin/c++, CXX_FLAGS= -Wno-deprecated -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -fopenmp -DNDEBUG -DUSE_KINETO -DUSE_FBGEMM -DUSE_QNNPACK -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DSYMBOLICATE_MOBILE_DEBUG_HANDLE -DEDGE_PROFILER_USE_KINETO -O2 -fPIC -Wno-narrowing -Wall -Wextra -Werror=return-type -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wno-sign-compare -Wno-unused-parameter -Wno-unused-variable -Wno-unused-function -Wno-unused-result -Wno-unused-local-typedefs -Wno-strict-overflow -Wno-strict-aliasing -Wno-error=deprecated-declarations -Wno-stringop-overflow -Wno-psabi -Wno-error=pedantic -Wno-error=redundant-decls -Wno-error=old-style-cast -fdiagnostics-color=always -faligned-new -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Wno-stringop-overflow, LAPACK_INFO=mkl, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, PERF_WITH_AVX512=1, TORCH_VERSION=1.10.0, USE_CUDA=ON, USE_CUDNN=ON, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_MKL=ON, USE_MKLDNN=ON, USE_MPI=OFF, USE_NCCL=ON, USE_NNPACK=ON, USE_OPENMP=ON,

TorchVision: 0.11.1
OpenCV: 4.7.0
MMCV: 1.4.0
MMCV Compiler: GCC 7.3
MMCV CUDA Compiler: 11.3
MMDetection: 2.13.0+c820f32

2023-09-05 21:09:00,423 - mmdet - INFO - Distributed training: False
2023-09-05 21:09:01,221 - mmdet - INFO - Config:
checkpoint_config = dict(interval=1)
log_config = dict(interval=50, hooks=[dict(type='TextLoggerHook')])
custom_hooks = [dict(type='NumClassCheckHook')]
dist_params = dict(backend='nccl')
log_level = 'INFO'
load_from = None
resume_from = None
workflow = [('train', 1)]
norm_cfg = dict(type='GN', num_groups=32, requires_grad=True)

Bug Description

[>>> ] 426/3822, 16.0 task/s, elapsed: 27s, ETA: 213s
[>>> ] 427/3822, 15.9 task/s, elapsed: 27s, ETA: 214s
[>>> ] 428/3822, 15.9 task/s, elapsed: 27s, ETA: 213s
[>>> ] 429/3822, 15.9 task/s, elapsed: 27s, ETA: 213s
[>>> ] 430/3822, 15.9 task/s, elapsed: 27s, ETA: 213s
[>>> ] 431/3822, 15.9 task/s, elapsed: 27s, ETA: 214s
[>>> ] 432/3822, 15.9 task/s, elapsed: 27s, ETA: 213s
[>>> ] 433/3822, 15.9 task/s, elapsed: 27s, ETA: 213s
[>>> ] 434/3822, 15.9 task/s, elapsed: 27s, ETA: 213sTraceback (most recent call last):
File "tools/train.py", line 192, in
main()
File "tools/train.py", line 181, in main
train_detector(
File "/media/dell/data1/ljw/code/test1/CPR/P2BNet/TOV_mmdetection/mmdet/apis/train.py", line 172, in train_detector
runner.run(data_loaders, cfg.workflow)
File "/media/dell/data1/miniconda3/envs/mmdetp2b/lib/python3.8/site-packages/mmcv/runner/epoch_based_runner.py", line 127, in run
epoch_runner(data_loaders[i], **kwargs)
File "/media/dell/data1/miniconda3/envs/mmdetp2b/lib/python3.8/site-packages/mmcv/runner/epoch_based_runner.py", line 54, in train
self.call_hook('after_train_epoch')
File "/media/dell/data1/miniconda3/envs/mmdetp2b/lib/python3.8/site-packages/mmcv/runner/base_runner.py", line 307, in call_hook
getattr(hook, fn_name)(self)
File "/media/dell/data1/miniconda3/envs/mmdetp2b/lib/python3.8/site-packages/mmcv/runner/hooks/evaluation.py", line 267, in after_train_epoch
self._do_evaluate(runner)
File "/media/dell/data1/ljw/code/test1/CPR/P2BNet/TOV_mmdetection/mmdet/core/evaluation/eval_hooks.py", line 44, in _do_evaluate
results = single_gpu_test(runner.model, self.dataloader, show=False)
File "/media/dell/data1/ljw/code/test1/CPR/P2BNet/TOV_mmdetection/mmdet/apis/test.py", line 27, in single_gpu_test
result = model(return_loss=False, rescale=True, **data)
File "/media/dell/data1/miniconda3/envs/mmdetp2b/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1102, in _call_impl
return forward_call(*input, **kwargs)
File "/media/dell/data1/miniconda3/envs/mmdetp2b/lib/python3.8/site-packages/mmcv/parallel/data_parallel.py", line 50, in forward
return super().forward(*inputs, **kwargs)
File "/media/dell/data1/miniconda3/envs/mmdetp2b/lib/python3.8/site-packages/torch/nn/parallel/data_parallel.py", line 166, in forward
return self.module(*inputs[0], **kwargs[0])
File "/media/dell/data1/miniconda3/envs/mmdetp2b/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1102, in _call_impl
return forward_call(*input, **kwargs)
File "/media/dell/data1/miniconda3/envs/mmdetp2b/lib/python3.8/site-packages/mmcv/runner/fp16_utils.py", line 98, in new_func
return old_func(*args, **kwargs)
File "/media/dell/data1/ljw/code/test1/CPR/P2BNet/TOV_mmdetection/mmdet/models/detectors/base.py", line 177, in forward
return self.forward_test(img, img_metas, **kwargs)
File "/media/dell/data1/ljw/code/test1/CPR/P2BNet/TOV_mmdetection/mmdet/models/detectors/base.py", line 150, in forward_test
return self.simple_test(imgs[0], img_metas[0], **kwargs)
File "/media/dell/data1/ljw/code/test1/CPR/P2BNet/TOV_mmdetection/mmdet/models/detectors/P2BNet.py", line 375, in simple_test
test_result, pseudo_boxes = self.roi_head.simple_test(stage,
File "/media/dell/data1/ljw/code/test1/CPR/P2BNet/TOV_mmdetection/mmdet/models/roi_heads/P2B_head.py", line 446, in simple_test
det_bboxes, det_labels, pseudo_bboxes = self.simple_test_bboxes(
File "/media/dell/data1/ljw/code/test1/CPR/P2BNet/TOV_mmdetection/mmdet/models/roi_heads/P2B_head.py", line 474, in simple_test_bboxes
bbox_results = self._bbox_forward(x, rois, gt_bboxes, stage)
File "/media/dell/data1/ljw/code/test1/CPR/P2BNet/TOV_mmdetection/mmdet/models/roi_heads/P2B_head.py", line 210, in _bbox_forward
bbox_feats = self.bbox_roi_extractor(
File "/media/dell/data1/miniconda3/envs/mmdetp2b/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1102, in call_impl
return forward_call(*input, **kwargs)
File "/media/dell/data1/miniconda3/envs/mmdetp2b/lib/python3.8/site-packages/mmcv/runner/fp16_utils.py", line 186, in new_func
return old_func(*args, **kwargs)
File "/media/dell/data1/ljw/code/test1/CPR/P2BNet/TOV_mmdetection/mmdet/models/roi_heads/roi_extractors/single_level_roi_extractor.py", line 102, in forward
roi_feats_t = self.roi_layers[i](feats[i], rois)
File "/media/dell/data1/miniconda3/envs/mmdetp2b/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1102, in _call_impl
return forward_call(*input, **kwargs)
File "/media/dell/data1/miniconda3/envs/mmdetp2b/lib/python3.8/site-packages/mmcv/ops/roi_align.py", line 212, in forward
return roi_align(input, rois, self.output_size, self.spatial_scale,
File "/media/dell/data1/miniconda3/envs/mmdetp2b/lib/python3.8/site-packages/mmcv/ops/roi_align.py", line 84, in forward
output = input.new_zeros(output_shape)
RuntimeError: CUDA out of memory. Tried to allocate 2.28 GiB (GPU 0; 23.65 GiB total capacity; 14.99 GiB already allocated; 1.28 GiB free; 21.04 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

Hello! What is the meaning of "off-the-shelf proposal"? As a freshman, I can't understand it. Sincerely, hope for your answer. (You can answer in Chinese or English, I can read Chinese. )

Hello, I would like to know if my data is originally marked with points annotations, but not marked with bbox, how to implement true box?

The meaning of gt_true_bbox and gt_bbox in the code

I set the batch size to 8 on one gpu, how should I set the learning rate
When I set the learning rate to 0.01, the ap value on coco2017 is only 20.0

Hi, i am very interested about the new method, i wonder if the baseline of this paper can be reproduced stably? Since when i use my computational resource to train this network, i find it hard to reach the same precision as in the paper. Thanks for your excellent work!

Dear authors:

Thanks for your great job! What are the mean-IoU results of P2BNet when only ResNet50 is used without the FPN help?

By the way, would it be helpful if I were to share the backbone of both P2B and FasterRCNN?

Hi, The google Google Drive link to download the pretrained weights doesn't work.

Hi! Thanks for your great job!
I am curious about the results of your proposed P2BNet on VOC2007. Could you please show me the performance? Thanks a lot!

I have tried to train on DOTA using the COCO config, but got a bad result, the minimum loss displayed when training was more than 0.1, and the pseudo box in visualization is terrible. Does this model not apply to remote sensing data sets? these are one pic of my results and train information:

The following code was executed

[cmd 0] inference with trained P2BNet to get pseudo box

work_dir='../TOV_mmdetection_cache/work_dir/coco/' && CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 PORT=10000 tools/dist_train.sh configs2/COCO/P2BNet/P2BNet_r50_fpn_1x_coco_ms.py 8
--work-dir=${work_dir}
--cfg-options evaluation.save_result_file=${work_dir}'_1200_latest_result.json' load_from=${work_dir}'P2BNet/epoch_12.pth' evaluation.do_first_eval=True runner.max_epochs=0

Wonder why the results are so bad

Hello, I would like to know how to generate point annotations in other data sets, such as voc2007

Can you update the link to the dataset and weights file, please?

Hello,I would like to know why the memory has been increasing when training?The device are 4* RTX3090.Is there a solution to solve the problem? Thank you.