CUDA out of memory error - 1920*1080 image - batch size is 1

[muito93]

I met error: Exception has occurred: RuntimeError CUDA out of memory. Tried to allocate 190.00 MiB (GPU 0; 3.94 GiB total capacity; 2.12 GiB already allocated; 171.06 MiB free; 2.22 GiB reserved in total by PyTorch) File "/media/thanhvt/HDD_Data/UbuntuWork/LenseProject/Lenses_Detectron2/train.py", line 121, in trainer.train()

I used custom Dataloaders. If I use T.RandomApply(transform=T.Resize(shape=(800, 800)), prob=1), it run well. But If I comment this line and use origin image size (1920*1080), CUDA out of memory appear. As I know, I use batch size is 1 (cfg.SOLVER.IMS_PER_BATCH = 1)

Instructions To Reproduce the Issue:

1. full code you wrote or full changes you made (git diff)

   
   import torch, torchvision
   import detectron2
   # Setup detectron2 logger
   from detectron2.utils.logger import setup_logger
   setup_logger()
   # import some common libraries
   import numpy as np
   import os, json, cv2, random
   # import some common detectron2 utilities
   from detectron2 import model_zoo
   from detectron2.engine import DefaultPredictor
   from detectron2.config import get_cfg
   from detectron2.utils.visualizer import Visualizer
   from detectron2.data import MetadataCatalog, DatasetCatalog,build_detection_test_loader,build_detection_train_loader

from detectron2.structures import BoxMode from detectron2.engine import DefaultTrainer from detectron2.data import DatasetMapper

def get_scratch_lense_dicts(img_dir): json_file = os.path.join(img_dir, "via_project_9Aug2020_9h36m_json.json") with open(json_file) as f: imgs_anns = json.load(f)

dataset_dicts = []
for idx, v in enumerate(imgs_anns.values()):
    record = {}

    filename = os.path.join(img_dir, v["filename"])
    height, width = cv2.imread(filename).shape[:2]

    record["file_name"] = filename
    record["image_id"] = idx
    record["height"] = height
    record["width"] = width

    annos = v["regions"]
    objs = []
    for anno in annos:
        assert not anno["region_attributes"]
        anno = anno["shape_attributes"]
        px = anno["x"]
        py = anno["y"]
        width = anno["width"]
        height = anno["height"]

        obj = {
            "bbox": [px, py, width, height],
            "bbox_mode": BoxMode.XYWH_ABS,
            "category_id": 0,
        }
        objs.append(obj)
    record["annotations"] = objs
    dataset_dicts.append(record)
return dataset_dicts

for d in ["train", "val"]: DatasetCatalog.register("scratchlense" + d, lambda d=d: get_scratch_lense_dicts("scratch_lense/" + d)) MetadataCatalog.get("scratchlense" + d).set(thing_classes=["scratch_lense"]) scratch_lense_metadata = MetadataCatalog.get("scratch_lense_train")

cfg = get_cfg() cfg.merge_from_file(model_zoo.get_config_file("COCO-Detection/faster_rcnn_R_50_FPN_3x.yaml")) cfg.DATASETS.TRAIN = ("scratch_lense_train",) cfg.DATASETS.TEST = () cfg.DATALOADER.NUM_WORKERS = 8 cfg.MODEL.WEIGHTS = model_zoo.get_checkpoint_url("COCO-Detection/faster_rcnn_R_50_FPN_3x.yaml") # Let training initialize from model zoo cfg.SOLVER.IMS_PER_BATCH = 1 cfg.SOLVER.BASE_LR = 0.00025 # pick a good LR cfg.SOLVER.MAX_ITER = 300 # 300 iterations seems good enough for this toy dataset; you may need to train longer for a practical dataset cfg.MODEL.ROI_HEADS.BATCH_SIZE_PER_IMAGE = 56 # faster, and good enough for this toy dataset (default: 512) cfg.MODEL.ROI_HEADS.NUM_CLASSES = 1 # only has one class (ballon) cfg.SOLVER.CHECKPOINT_PERIOD = 50 cfg.MAX_SIZE_TRAIN = 2000

os.makedirs(cfg.OUTPUT_DIR, exist_ok=True)

import detectron2.data.transforms as T from detectron2.data import detection_utils as utils import copy def custom_mapper(dataset_dict):

Implement a mapper, similar to the default DatasetMapper, but with your own customizations

dataset_dict = copy.deepcopy(dataset_dict)  # it will be modified by code below
image = utils.read_image(dataset_dict["file_name"], format="BGR")

image, transforms = T.apply_transform_gens([
    T.RandomFlip(prob=0.5, horizontal=True, vertical=False),
    T.RandomFlip(prob=0.5, horizontal=False, vertical=True),
    # T.RandomApply(transform=T.Resize(shape=(800, 800)), 
    #               prob=1),### If I uncomment this line, there is CUDA memory error
    T.RandomApply(transform=T.RandomRotation(angle=[-45,45], expand=True, center=None, sample_style="range", interp=None), 
                  prob=0.5)

], image)

dataset_dict["image"] = torch.as_tensor(image.transpose(2, 0, 1).astype("float32"))

annos = [
    utils.transform_instance_annotations(obj, transforms, image.shape[:2])
    for obj in dataset_dict.pop("annotations")
    if obj.get("iscrowd", 0) == 0
]
instances = utils.annotations_to_instances(annos, image.shape[:2])
dataset_dict["instances"] = utils.filter_empty_instances(instances)
return dataset_dict

class CS_Trainer(DefaultTrainer): @classmethod def build_test_loader(cls, cfg, dataset_name): return build_detection_test_loader(cfg, dataset_name, mapper=DatasetMapper(cfg, False))

@classmethod
def build_train_loader(cls, cfg):
    return build_detection_train_loader(cfg, mapper=custom_mapper)

trainer = CS_Trainer(cfg) trainer.resume_or_load(resume=False) trainer.train()

from detectron2.evaluation import COCOEvaluator, inference_on_dataset from detectron2.data import build_detection_test_loader evaluator = COCOEvaluator("scratch_lense_val", cfg, False, output_dir="./output/") val_loader = build_detection_test_loader(cfg, "scratch_lense_val") print(inference_on_dataset(trainer.model, val_loader, evaluator))

2. what exact command you run:
3. __full logs__ you observed:

        (norm): FrozenBatchNorm2d(num_features=128, eps=1e-05)
      )
      (conv3): Conv2d(
        128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=512, eps=1e-05)
      )
    )
    (2): BottleneckBlock(
      (conv1): Conv2d(
        512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=128, eps=1e-05)
      )
      (conv2): Conv2d(
        128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=128, eps=1e-05)
      )
      (conv3): Conv2d(
        128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=512, eps=1e-05)
      )
    )
    (3): BottleneckBlock(
      (conv1): Conv2d(
        512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=128, eps=1e-05)
      )
      (conv2): Conv2d(
        128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=128, eps=1e-05)
      )
      (conv3): Conv2d(
        128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=512, eps=1e-05)
      )
    )
  )
  (res4): Sequential(
    (0): BottleneckBlock(
      (shortcut): Conv2d(
        512, 1024, kernel_size=(1, 1), stride=(2, 2), bias=False
        (norm): FrozenBatchNorm2d(num_features=1024, eps=1e-05)
      )
      (conv1): Conv2d(
        512, 256, kernel_size=(1, 1), stride=(2, 2), bias=False
        (norm): FrozenBatchNorm2d(num_features=256, eps=1e-05)
      )
      (conv2): Conv2d(
        256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=256, eps=1e-05)
      )
      (conv3): Conv2d(
        256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=1024, eps=1e-05)
      )
    )
    (1): BottleneckBlock(
      (conv1): Conv2d(
        1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=256, eps=1e-05)
      )
      (conv2): Conv2d(
        256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=256, eps=1e-05)
      )
      (conv3): Conv2d(
        256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=1024, eps=1e-05)
      )
    )
    (2): BottleneckBlock(
      (conv1): Conv2d(
        1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=256, eps=1e-05)
      )
      (conv2): Conv2d(
        256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=256, eps=1e-05)
      )
      (conv3): Conv2d(
        256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=1024, eps=1e-05)
      )
    )
    (3): BottleneckBlock(
      (conv1): Conv2d(
        1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=256, eps=1e-05)
      )
      (conv2): Conv2d(
        256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=256, eps=1e-05)
      )
      (conv3): Conv2d(
        256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=1024, eps=1e-05)
      )
    )
    (4): BottleneckBlock(
      (conv1): Conv2d(
        1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=256, eps=1e-05)
      )
      (conv2): Conv2d(
        256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=256, eps=1e-05)
      )
      (conv3): Conv2d(
        256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=1024, eps=1e-05)
      )
    )
    (5): BottleneckBlock(
      (conv1): Conv2d(
        1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=256, eps=1e-05)
      )
      (conv2): Conv2d(
        256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=256, eps=1e-05)
      )
      (conv3): Conv2d(
        256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=1024, eps=1e-05)
      )
    )
  )
  (res5): Sequential(
    (0): BottleneckBlock(
      (shortcut): Conv2d(
        1024, 2048, kernel_size=(1, 1), stride=(2, 2), bias=False
        (norm): FrozenBatchNorm2d(num_features=2048, eps=1e-05)
      )
      (conv1): Conv2d(
        1024, 512, kernel_size=(1, 1), stride=(2, 2), bias=False
        (norm): FrozenBatchNorm2d(num_features=512, eps=1e-05)
      )
      (conv2): Conv2d(
        512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=512, eps=1e-05)
      )
      (conv3): Conv2d(
        512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=2048, eps=1e-05)
      )
    )
    (1): BottleneckBlock(
      (conv1): Conv2d(
        2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=512, eps=1e-05)
      )
      (conv2): Conv2d(
        512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=512, eps=1e-05)
      )
      (conv3): Conv2d(
        512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=2048, eps=1e-05)
      )
    )
    (2): BottleneckBlock(
      (conv1): Conv2d(
        2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=512, eps=1e-05)
      )
      (conv2): Conv2d(
        512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=512, eps=1e-05)
      )
      (conv3): Conv2d(
        512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False
        (norm): FrozenBatchNorm2d(num_features=2048, eps=1e-05)
      )
    )
  )
)

) (proposal_generator): RPN( (rpn_head): StandardRPNHead( (conv): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) (objectness_logits): Conv2d(256, 3, kernel_size=(1, 1), stride=(1, 1)) (anchor_deltas): Conv2d(256, 12, kernel_size=(1, 1), stride=(1, 1)) ) (anchor_generator): DefaultAnchorGenerator( (cell_anchors): BufferList() ) ) (roi_heads): StandardROIHeads( (box_pooler): ROIPooler( (level_poolers): ModuleList( (0): ROIAlign(output_size=(7, 7), spatial_scale=0.25, sampling_ratio=0, aligned=True) (1): ROIAlign(output_size=(7, 7), spatial_scale=0.125, sampling_ratio=0, aligned=True) (2): ROIAlign(output_size=(7, 7), spatial_scale=0.0625, sampling_ratio=0, aligned=True) (3): ROIAlign(output_size=(7, 7), spatial_scale=0.03125, sampling_ratio=0, aligned=True) ) ) (box_head): FastRCNNConvFCHead( (fc1): Linear(in_features=12544, out_features=1024, bias=True) (fc2): Linear(in_features=1024, out_features=1024, bias=True) ) (box_predictor): FastRCNNOutputLayers( (cls_score): Linear(in_features=1024, out_features=2, bias=True) (bbox_pred): Linear(in_features=1024, out_features=4, bias=True) ) ) ) [08/14 07:32:20 d2.data.build]: Removed 0 images with no usable annotations. 17 images left. [08/14 07:32:20 d2.data.build]: Distribution of instances among all 1 categories:	category	#instances
scratch_lense	17

[08/14 07:32:20 d2.data.common]: Serializing 17 elements to byte tensors and concatenating them all ... [08/14 07:32:20 d2.data.common]: Serialized dataset takes 0.00 MiB [08/14 07:32:20 d2.data.build]: Using training sampler TrainingSampler 2020-08-14 07:32:20.970958: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcudart.so.10.1 Skip loading parameter 'roi_heads.box_predictor.cls_score.weight' to the model due to incompatible shapes: (81, 1024) in the checkpoint but (2, 1024) in the model! You might want to double check if this is expected. Skip loading parameter 'roi_heads.box_predictor.cls_score.bias' to the model due to incompatible shapes: (81,) in the checkpoint but (2,) in the model! You might want to double check if this is expected. Skip loading parameter 'roi_heads.box_predictor.bbox_pred.weight' to the model due to incompatible shapes: (320, 1024) in the checkpoint but (4, 1024) in the model! You might want to double check if this is expected. Skip loading parameter 'roi_heads.box_predictor.bbox_pred.bias' to the model due to incompatible shapes: (320,) in the checkpoint but (4,) in the model! You might want to double check if this is expected. [08/14 07:32:23 d2.engine.train_loop]: Starting training from iteration 0 /home/thanhvt/anaconda3/envs/keras/lib/python3.8/site-packages/detectron2/layers/wrappers.py:226: UserWarning: This overload of nonzero is deprecated: nonzero() Consider using one of the following signatures instead: nonzero(, bool as_tuple) (Triggered internally at /pytorch/torch/csrc/utils/python_arg_parser.cpp:766.) return x.nonzero().unbind(1) ERROR [08/14 07:32:25 d2.engine.train_loop]: Exception during training: Traceback (most recent call last): File "/home/thanhvt/anaconda3/envs/keras/lib/python3.8/site-packages/detectron2/engine/train_loop.py", line 140, in train self.run_step() File "/home/thanhvt/anaconda3/envs/keras/lib/python3.8/site-packages/detectron2/engine/train_loop.py", line 226, in run_step loss_dict = self.model(data) File "/home/thanhvt/anaconda3/envs/keras/lib/python3.8/site-packages/torch/nn/modules/module.py", line 722, in _call_impl result = self.forward(input, kwargs) File "/home/thanhvt/anaconda3/envs/keras/lib/python3.8/site-packages/detectron2/modeling/meta_arch/rcnn.py", line 157, in forward features = self.backbone(images.tensor) File "/home/thanhvt/anaconda3/envs/keras/lib/python3.8/site-packages/torch/nn/modules/module.py", line 722, in _call_impl result = self.forward(*input, *kwargs) File "/home/thanhvt/anaconda3/envs/keras/lib/python3.8/site-packages/detectron2/modeling/backbone/fpn.py", line 132, in forward lateral_features = lateral_conv(features) File "/home/thanhvt/anaconda3/envs/keras/lib/python3.8/site-packages/torch/nn/modules/module.py", line 722, in _call_impl result = self.forward(input, kwargs) File "/home/thanhvt/anaconda3/envs/keras/lib/python3.8/site-packages/detectron2/layers/wrappers.py", line 94, in forward x = super().forward(x) File "/home/thanhvt/anaconda3/envs/keras/lib/python3.8/site-packages/torch/nn/modules/conv.py", line 419, in forward return self._conv_forward(input, self.weight) File "/home/thanhvt/anaconda3/envs/keras/lib/python3.8/site-packages/torch/nn/modules/conv.py", line 415, in _conv_forward return F.conv2d(input, weight, self.bias, self.stride, RuntimeError: CUDA out of memory. Tried to allocate 190.00 MiB (GPU 0; 3.94 GiB total capacity; 2.12 GiB already allocated; 171.06 MiB free; 2.22 GiB reserved in total by PyTorch) [08/14 07:32:25 d2.engine.hooks]: Total training time: 0:00:01 (0:00:00 on hooks)

## Expected behavior:

If there are no obvious error in "what you observed" provided above,
please tell us the expected behavior.

If you expect the model to converge / work better, note that we do not give suggestions
on how to train a new model.
Only in one of the two conditions we will help with it:
(1) You're unable to reproduce the results in detectron2 model zoo.
(2) It indicates a detectron2 bug.

## Environment:

Provide your environment information using the following command:

---

sys.platform linux Python 3.8.3 (default, Jul 2 2020, 16:21:59) [GCC 7.3.0] numpy 1.18.5 detectron2 0.2.1 @/home/thanhvt/anaconda3/envs/keras/lib/python3.8/site-packages/detectron2 Compiler GCC 7.3 CUDA compiler CUDA 10.1 detectron2 arch flags sm_35, sm_37, sm_50, sm_52, sm_60, sm_61, sm_70, sm_75 DETECTRON2_ENV_MODULE PyTorch 1.6.0+cu101 @/home/thanhvt/anaconda3/envs/keras/lib/python3.8/site-packages/torch PyTorch debug build False GPU available True GPU 0 GeForce GTX 1050 Ti CUDA_HOME /usr/local/cuda-10.1 Pillow 7.2.0 torchvision 0.7.0+cu101 @/home/thanhvt/anaconda3/envs/keras/lib/python3.8/site-packages/torchvision torchvision arch flags sm_35, sm_50, sm_60, sm_70, sm_75 fvcore 0.1.1.post20200716 cv2 4.3.0

---

PyTorch built with:

• GCC 7.3
• C++ Version: 201402
• Intel(R) Math Kernel Library Version 2019.0.5 Product Build 20190808 for Intel(R) 64 architecture applications
• Intel(R) MKL-DNN v1.5.0 (Git Hash e2ac1fac44c5078ca927cb9b90e1b3066a0b2ed0)
• OpenMP 201511 (a.k.a. OpenMP 4.5)
• NNPACK is enabled
• CPU capability usage: AVX2
• CUDA Runtime 10.1
• 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_70,code=sm_70;-gencode;arch=compute_75,code=sm_75
• CuDNN 7.6.3
• Magma 2.5.2
• Build settings: BLAS=MKL, BUILD_TYPE=Release, CXX_FLAGS= -Wno-deprecated -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -fopenmp -DNDEBUG -DUSE_FBGEMM -DUSE_QNNPACK -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DUSE_VULKAN_WRAPPER -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-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, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, PERF_WITH_AVX512=1, USE_CUDA=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, USE_STATIC_DISPATCH=OFF,

[ppwwyyxx]

error is expected given the small total memory size.

[pvti]

This also happens in another machine with Nvidia GTX 1080 Ti 11Gb GPU memory. So how should we handle this typical case? We really need to use Custom Dataloader image transformation (withoud resizing the original image to 800x800 as in the Default Dataloader) due to information loss. @ppwwyyxx Any suggestion would be appriciated. Thanks.

[ppwwyyxx]

Suggestions are to use smaller images, smaller models or larger GPUs.

[alexriedel1]

Your random rotation augementation might also be a problem as it increases memory allocation