deeplab-ai / atr-net Goto Github PK

I have been trying to run your code for sometime. But I was having error while loading the dataset. I changed line 34 of src/utils/file_utils.py as follows.

34: orig_img_names = set(os.listdir(ORIG_IMAGES_PATH[dataset]))

Previously it was ORIG_ANNOS_PATHS and the annotations variables was being null.

Now the training is going on. You might want to look into this issue.

Hello, I have installed and setup the code base in my system. However, I have run into an issue and I am having difficulty resolving it. I believe it has something to do with the version mismatch of the different libraries, but I couldn't pin-point the problem. Here is the Error Message:

File "/content/drive/My Drive/DirResearch/atr-net/faster_rcnn/model/faster_rcnn/faster_rcnn.py", line 10, in <module>
    from model.rpn.rpn import _RPN
  File "/content/drive/My Drive/DirResearch/atr-net/faster_rcnn/model/rpn/rpn.py", line 8, in <module>
    from .proposal_layer import _ProposalLayer
  File "/content/drive/My Drive/DirResearch/atr-net/faster_rcnn/model/rpn/proposal_layer.py", line 20, in <module>
    from model.nms.nms_wrapper import nms
  File "/content/drive/My Drive/DirResearch/atr-net/faster_rcnn/model/nms/nms_wrapper.py", line 10, in <module>
    from model.nms.nms_gpu import nms_gpu
  File "/content/drive/My Drive/DirResearch/atr-net/faster_rcnn/model/nms/nms_gpu.py", line 4, in <module>
    from ._ext import nms
  File "/content/drive/My Drive/DirResearch/atr-net/faster_rcnn/model/nms/_ext/nms/__init__.py", line 2, in <module>
    from torch.utils.ffi import _wrap_function
  File "/usr/local/lib/python3.6/dist-packages/torch/utils/ffi/__init__.py", line 1, in <module>
    raise ImportError("torch.utils.ffi is deprecated. Please use cpp extensions instead.")
ImportError: torch.utils.ffi is deprecated. Please use cpp extensions instead.

Please help me resolve this error.

Thank you

Hello.

Where we are getting this anno dictionary in file_utils.py? I am a bit confused about the use here.

Dear scholar,
I want to ask about the size of cuda memory by running the code

python3 main.py --dataset=VG200 --task=sggen --model=atr_net

I want to evaluate the model. And my cuda is 24G but running the below code will throw the error "RuntimeError: CUDA error: out of memory"

/home/byd/atr-net/faster_rcnn/model/utils/config.py:375: YAMLLoadWarning: calling yaml.load() without Loader=... is deprecated, as the default Loader is unsafe. Please read https://msg.pyyaml.org/load for full details.
yaml_cfg = edict(yaml.load(f))
load checkpoint faster_rcnn/faster_rcnn_1_10_14657.pth
Traceback (most recent call last):
File "main.py", line 8, in
from src.models import (
File "/home/byd/atr-net/src/models/atr_net.py", line 21, in
from src.utils.train_test_utils import SGGTrainTester
File "/home/byd/atr-net/src/utils/train_test_utils.py", line 13, in
from src.utils.data_loading_utils import (
File "/home/byd/atr-net/src/utils/data_loading_utils.py", line 14, in
FEATURE_EXTRACTOR = BaseFeatureExtractor()
File "/home/byd/atr-net/src/tools/feature_extractors.py", line 80, in init
checkpoint = _load()
File "/home/byd/atr-net/src/tools/feature_extractors.py", line 67, in _load
checkpoint = torch.load(load_name)
File "/home/byd/.conda/envs/vilbert-mt/lib/python3.6/site-packages/torch/serialization.py", line 593, in load
return _legacy_load(opened_file, map_location, pickle_module, **pickle_load_args)
File "/home/byd/.conda/envs/vilbert-mt/lib/python3.6/site-packages/torch/serialization.py", line 772, in _legacy_load
result = unpickler.load()
File "/home/byd/.conda/envs/vilbert-mt/lib/python3.6/site-packages/torch/serialization.py", line 728, in persistent_load
deserialized_objects[root_key] = restore_location(obj, location)
File "/home/byd/.conda/envs/vilbert-mt/lib/python3.6/site-packages/torch/serialization.py", line 175, in default_restore_location
result = fn(storage, location)
File "/home/byd/.conda/envs/vilbert-mt/lib/python3.6/site-packages/torch/serialization.py", line 155, in _cuda_deserialize
return storage_type(obj.size())
File "/home/byd/.conda/envs/vilbert-mt/lib/python3.6/site-packages/torch/cuda/init.py", line 484, in _lazy_new
return super(_CudaBase, cls).new(cls, *args, **kwargs)
RuntimeError: CUDA error: out of memory

I would like to use your model to process a set of raw images into scene graphs. It is unclear to me which tasks I have to pipeline in order to process a raw image all the way to a scene graph.

What are the steps to pretrain on VG200 and then run scene graph generation on a new set of images?

I have tried checking the config and data_loading_utils file, but still unable to fix this error.
From the error,k it's not reading the dataset as the len definition if data_loading_utils returning zero.
@nickgkan Could you please help to fix this error? Thanks!

/home/kgl/atr-net/faster_rcnn/model/utils/config.py:374: YAMLLoadWarning: calling yaml.load() without Loader=... is deprecated, as the default Loader is unsafe. Please read https://msg.pyyaml.org/load for full details.
  yaml_cfg = edict(yaml.load(f))
load checkpoint faster_rcnn/faster_rcnn_1_10_14657.pth
2020-01-14 11:58:45,650 - DEBUG - Tackling predcls for 1 classes
{'attention': 'multi_head', 'use_language': True, 'use_spatial': True}
load checkpoint faster_rcnn/faster_rcnn_1_10_14657.pth
2020-01-14 11:58:46,947 - INFO - Performing training for atr_net_predcls_VG80K
2020-01-14 11:58:47,904 - DEBUG - Set up dataset of 0 files
2020-01-14 11:58:47,904 - DEBUG - Set up dataset of 0 files
2020-01-14 11:58:47,904 - DEBUG - Set up dataset of 0 files
Traceback (most recent call last):
  File "main.py", line 113, in <module>
    main()
  File "main.py", line 110, in main
    model.train_test(cfg)
  File "/home/kgl/atr-net/src/models/atr_net.py", line 487, in train_test
    epochs=30 if config.use_early_stopping else 7)
  File "/home/kgl/atr-net/src/utils/train_test_utils.py", line 81, in train
    self._set_data_loaders()
  File "/home/kgl/atr-net/src/utils/train_test_utils.py", line 358, in _set_data_loaders
    for split in mode_ids
  File "/home/kgl/atr-net/src/utils/train_test_utils.py", line 358, in <dictcomp>
    for split in mode_ids
  File "/home/kgl/atr-net/src/utils/data_loading_utils.py", line 320, in __init__
    collate_fn=collate_fn)
  File "/home/kgl/.local/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 802, in __init__
    sampler = RandomSampler(dataset)
  File "/home/kgl/.local/lib/python3.6/site-packages/torch/utils/data/sampler.py", line 64, in __init__
    "value, but got num_samples={}".format(self.num_samples))
ValueError: num_samples should be a positive integeral value, but got num_samples=0

First, thanks for your code. I want to train a model use the code as follow:

python3 main.py --dataset=VG200 --task=sggen --model=atr_net.

But, I got the error:

load checkpoint faster_rcnn/faster_rcnn_1_10_14657.pth
Tackling sggen for 51 classes
{'attention': 'multi_head', 'use_language': True, 'use_spatial': True}
2021-07-08 21:47:58,190 - DEBUG - Tackling sggen for 51 classes
load checkpoint faster_rcnn/faster_rcnn_1_10_14657.pth
2021-07-08 21:48:00,201 - INFO - Performing training for atr_net_predcls_VG200
Performing training for atr_net_predcls_VG200
Traceback (most recent call last):
File "/home/nimo/code_sgg/atr-net/main.py", line 113, in
main()
File "/home/nimo/code_sgg/atr-net/main.py", line 110, in main
model.train_test(cfg)
File "/mnt/e/wsl_code/atr-net/src/models/atr_net.py", line 488, in train_test
epochs=30 if config.use_early_stopping else 7)
File "/mnt/e/wsl_code/atr-net/src/utils/train_test_utils.py", line 82, in train
self._set_data_loaders()
File "/mnt/e/wsl_code/atr-net/src/utils/train_test_utils.py", line 334, in _set_data_loaders
self._filter_duplicate_rels, self.filter_multiple_preds)[0])
File "/mnt/e/wsl_code/atr-net/src/utils/file_utils.py", line 31, in load_annotations
with open(PATHS['json_path'] + dataset + '' + _mode + '.json') as fid:
FileNotFoundError: [Errno 2] No such file or directory: 'json_annos/VG200_sggen.json'

Process finished with exit code 1

I check json_annos dir, but I didn't find the *_sggen.json. I also check the code of prepare_data.py and ***_transformer_class.py , howevery there are not any information for **sggen.json.

Can you help me?

Thank you very much!