TypeError when training

[TNT028]

Thanks for you code. However, when I run "CUDA_VISIBLE_DEVICES=0 python -m torch.distributed.launch --master_port 10025 --nproc_per_node=1 tools/relation_train_net.py --config-file "configs/SHA_GCL_e2e_relation_X_101_32_8_FPN_1x.yaml" GLOBAL_SETTING.DATASET_CHOICE 'VG' GLOBAL_SETTING.RELATION_PREDICTOR 'TransLike_GCL' GLOBAL_SETTING.BASIC_ENCODER 'Hybrid-Attention' GLOBAL_SETTING.GCL_SETTING.GROUP_SPLIT_MODE 'divide4' GLOBAL_SETTING.GCL_SETTING.KNOWLEDGE_TRANSFER_MODE 'KL_logit_TopDown' MODEL.ROI_RELATION_HEAD.USE_GT_BOX True MODEL.ROI_RELATION_HEAD.USE_GT_OBJECT_LABEL True SOLVER.IMS_PER_BATCH 8 TEST.IMS_PER_BATCH 8 DTYPE "float16" SOLVER.MAX_ITER 60000 SOLVER.VAL_PERIOD 5000 SOLVER.CHECKPOINT_PERIOD 5000 GLOVE_DIR /home/yons/scene/VTSCN/datasets/vg/glove OUTPUT_DIR /home/yons/scene/VTSCN/datasets/output/SHA_GCL_VG_PredCls_test" with SHA-GCL-for-SGG for training PredCls, it produces the following issue: TypeError:init() takes 3 positional arguments but 5 were given.

Hope to your reply. Thank you!

There is the whole issue txt:

Traceback (most recent call last): File "tools/relation_train_net.py", line 442, in main() File "tools/relation_train_net.py", line 435, in main model = train(cfg, args.local_rank, args.distributed, logger) File "tools/relation_train_net.py", line 48, in train model = build_detection_model(cfg) File "/home/yons/scene/VTSCN/SHA-GCL-for-SGG/maskrcnn_benchmark/modeling/detector/detectors.py", line 10, in build_detection_model return meta_arch(cfg) File "/home/yons/scene/VTSCN/SHA-GCL-for-SGG/maskrcnn_benchmark/modeling/detector/generalized_rcnn.py", line 31, in init self.roi_heads = build_roi_heads(cfg, self.backbone.out_channels) File "/home/yons/scene/VTSCN/SHA-GCL-for-SGG/maskrcnn_benchmark/modeling/roi_heads/roi_heads.py", line 83, in build_roi_heads roi_heads.append(("box", build_roi_box_head(cfg, in_channels))) File "/home/yons/scene/VTSCN/SHA-GCL-for-SGG/maskrcnn_benchmark/modeling/roi_heads/box_head/box_head.py", line 115, in build_roi_box_head return ROIBoxHead(cfg, in_channels) File "/home/yons/scene/VTSCN/SHA-GCL-for-SGG/maskrcnn_benchmark/modeling/roi_heads/box_head/box_head.py", line 26, in init self.feature_extractor = make_roi_box_feature_extractor(cfg, in_channels, half_out=self.cfg.MODEL.ATTRIBUTE_ON) File "/home/yons/scene/VTSCN/SHA-GCL-for-SGG/maskrcnn_benchmark/modeling/roi_heads/box_head/roi_box_feature_extractors.py", line 166, in make_roi_box_feature_extractor return func(cfg, in_channels, half_out, cat_all_levels) TypeError: init() takes 3 positional arguments but 5 were given Traceback (most recent call last): File "/home/yons/anaconda3/envs/VTSCN/lib/python3.8/runpy.py", line 194, in _run_module_as_main return _run_code(code, main_globals, None, File "/home/yons/anaconda3/envs/VTSCN/lib/python3.8/runpy.py", line 87, in _run_code exec(code, run_globals) File "/home/yons/anaconda3/envs/VTSCN/lib/python3.8/site-packages/torch/distributed/launch.py", line 260, in main() File "/home/yons/anaconda3/envs/VTSCN/lib/python3.8/site-packages/torch/distributed/launch.py", line 255, in main raise subprocess.CalledProcessError(returncode=process.returncode, subprocess.CalledProcessError: Command '['/home/yons/anaconda3/envs/VTSCN/bin/python', '-u', 'tools/relation_train_net.py', '--local_rank=0', '--config-file', 'configs/SHA_GCL_e2e_relation_X_101_32_8_FPN_1x.yaml', 'GLOBAL_SETTING.DATASET_CHOICE', 'VG', 'GLOBAL_SETTING.RELATION_PREDICTOR', 'TransLike_GCL', 'GLOBAL_SETTING.BASIC_ENCODER', 'Hybrid-Attention', 'GLOBAL_SETTING.GCL_SETTING.GROUP_SPLIT_MODE', 'divide4', 'GLOBAL_SETTING.GCL_SETTING.KNOWLEDGE_TRANSFER_MODE', 'KL_logit_TopDown', 'MODEL.ROI_RELATION_HEAD.USE_GT_BOX', 'True', 'MODEL.ROI_RELATION_HEAD.USE_GT_OBJECT_LABEL', 'True', 'SOLVER.IMS_PER_BATCH', '8', 'TEST.IMS_PER_BATCH', '8', 'DTYPE', 'float16', 'SOLVER.MAX_ITER', '60000', 'SOLVER.VAL_PERIOD', '5000', 'SOLVER.CHECKPOINT_PERIOD', '5000', 'GLOVE_DIR', '/home/yons/scene/VTSCN/datasets/vg/glove', 'OUTPUT_DIR', '/home/yons/scene/VTSCN/datasets/output/SHA_GCL_VG_PredCls_test']' returned non-zero exit status 1.

[Nora-Zhang98]

Thanks for your recognition of our work. maskrcnn_benchmark/modeling/roi_heads/box_head/box_head.py is uploaded with a new version, you can download it and try again. BTW, I will delete the implementation of VTSCN of TransLike_GCL because they are incomplete and has not been run before. In our paper, we reported three baseline models, i.e. MOTIFS, VCTree and VTransE.

[TNT028]

Thanks for your recognition of our work. maskrcnn_benchmark/modeling/roi_heads/box_head/box_head.py is uploaded with a new version, you can download it and try again. The occurence of this error is because some codes in maskrcnn_benchmark are from the original benchmark of TDE. The dismatch between SHA-GCL and TDE causes this error. I am not sure if there are another dismatch errors. If you meet again, please let me know and I will recheck the code. BTW, I will delete the implementation of VTSCN of TransLike_GCL because they are incomplete and has not been run before. In our paper, we reported three baseline models, i.e. MOTIFS, VCTree and VTransE.

Thanks for your reply. I am sorry to bother you again. I have tried this new py and this issue is solved. But a new problem has arisen.

Traceback (most recent call last): File "tools/relation_train_net.py", line 24, in from maskrcnn_benchmark.modeling.detector import build_detection_model File "/home/yons/scene/VTSCN/SHA-GCL-for-SGG/maskrcnn_benchmark/modeling/detector/init.py", line 2, in from .detectors import build_detection_model File "/home/yons/scene/VTSCN/SHA-GCL-for-SGG/maskrcnn_benchmark/modeling/detector/detectors.py", line 2, in from .generalized_rcnn import GeneralizedRCNN File "/home/yons/scene/VTSCN/SHA-GCL-for-SGG/maskrcnn_benchmark/modeling/detector/generalized_rcnn.py", line 13, in from ..roi_heads.roi_heads import build_roi_heads File "/home/yons/scene/VTSCN/SHA-GCL-for-SGG/maskrcnn_benchmark/modeling/roi_heads/roi_heads.py", line 4, in from .box_head.box_head import build_roi_box_head ImportError: cannot import name 'build_roi_box_head' from 'maskrcnn_benchmark.modeling.roi_heads.box_head.box_head' (/home/yons/scene/VTSCN/SHA-GCL-for-SGG/maskrcnn_benchmark/modeling/roi_heads/box_head/box_head.py) Traceback (most recent call last): File "/home/yons/anaconda3/envs/VTSCN/lib/python3.8/runpy.py", line 194, in _run_module_as_main return _run_code(code, main_globals, None, File "/home/yons/anaconda3/envs/VTSCN/lib/python3.8/runpy.py", line 87, in _run_code exec(code, run_globals) File "/home/yons/anaconda3/envs/VTSCN/lib/python3.8/site-packages/torch/distributed/launch.py", line 260, in main() File "/home/yons/anaconda3/envs/VTSCN/lib/python3.8/site-packages/torch/distributed/launch.py", line 255, in main raise subprocess.CalledProcessError(returncode=process.returncode, subprocess.CalledProcessError: Command '['/home/yons/anaconda3/envs/VTSCN/bin/python', '-u', 'tools/relation_train_net.py', '--local_rank=0', '--config-file', 'configs/SHA_GCL_e2e_relation_X_101_32_8_FPN_1x.yaml', 'GLOBAL_SETTING.DATASET_CHOICE', 'VG', 'GLOBAL_SETTING.RELATION_PREDICTOR', 'MotifsLikePredictor', 'GLOBAL_SETTING.BASIC_ENCODER', 'Motifs', 'GLOBAL_SETTING.GCL_SETTING.GROUP_SPLIT_MODE', 'divide4', 'GLOBAL_SETTING.GCL_SETTING.KNOWLEDGE_TRANSFER_MODE', 'KL_logit_TopDown', 'MODEL.ROI_RELATION_HEAD.USE_GT_BOX', 'True', 'MODEL.ROI_RELATION_HEAD.USE_GT_OBJECT_LABEL', 'True', 'SOLVER.IMS_PER_BATCH', '8', 'TEST.IMS_PER_BATCH', '8', 'DTYPE', 'float16', 'SOLVER.MAX_ITER', '60000', 'SOLVER.VAL_PERIOD', '5000', 'SOLVER.CHECKPOINT_PERIOD', '5000', 'GLOVE_DIR', '/home/yons/scene/VTSCN/datasets/vg/glove', 'OUTPUT_DIR', '/home/yons/scene/VTSCN/datasets/output/SHA_GCL_VG_PredCls_test']' returned non-zero exit status 1.

[Nora-Zhang98]

Oh, the wrong file is maskrcnn_benchmark/modeling/roi_heads/box_head/roi_box_feature_extractors.py, but I mix it up with maskrcnn_benchmark/modeling/roi_heads/box_head/box_head.py. I am sorry, and please download these two files again. If you want to reimplement our VTSCN, 'GLOBAL_SETTING.RELATION_PREDICTOR' needs to be 'MotifsLike_GCL' or 'VCTree_GCL'.

[Nora-Zhang98]

But the first bug you raised may occur again because the codes are not changed actually. As the first bug shown that File "/home/yons/scene/VTSCN/SHA-GCL-for-SGG/maskrcnn_benchmark/modeling/roi_heads/box_head/roi_box_feature_extractors.py", line 166, in make_roi_box_feature_extractor return func(cfg, in_channels, half_out, cat_all_levels) TypeError: init() takes 3 positional arguments but 5 were given The ROI_BOX_HEAD.FEATURE_EXTRACTOR is FPN2MLPFeatureExtractor by default, which should take 5 arguments(roi_box_feature_extractor.py, line55): def init(self, cfg, in_channels, half_out=False, cat_all_levels=False): Maybe you can debug and check if the ROI_BOX_HEAD.FEATURE_EXTRACTOR is set correctly.

[TNT028]

But the first bug you raised may occur again because the codes are not changed actually. As the first bug shown that File "/home/yons/scene/VTSCN/SHA-GCL-for-SGG/maskrcnn_benchmark/modeling/roi_heads/box_head/roi_box_feature_extractors.py", line 166, in make_roi_box_feature_extractor return func(cfg, in_channels, half_out, cat_all_levels) TypeError: init() takes 3 positional arguments but 5 were given The ROI_BOX_HEAD.FEATURE_EXTRACTOR is FPN2MLPFeatureExtractor by default, which should take 5 arguments(roi_box_feature_extractor.py, line55): def init(self, cfg, in_channels, half_out=False, cat_all_levels=False): Maybe you can debug and check if the ROI_BOX_HEAD.FEATURE_EXTRACTOR is set correctly.

You are right. It has indeed reappeared again. Thank you for your guidance. I'll give it a try to check. If you have any new ideas, please contact me as soon as possible. I would greatly appreciate it.

[Nora-Zhang98]

I have downloaded my code and rerun them, and no bug occurs. The GLOBAL_SETTING.RELATION_PREDICTOR is set as 'MotifsLike_GCL', and GLOBAL_SETTING.BASIC_ENCODER is set as 'Motifs'.