Hi,
I want retrain faster rcnn modl from tlt, and I set resume_from_model: “/workspace/tlt-experiments/data/faster_rcnn/frcnn_kitti_darknet53.epoch111.tlt”
However OOM error happend. I dont think the GPU memory is not enough, I have set batch_size_per_gpu 1 and export TF_FORCE_GPU_ALLOW_GROWTH=true. Meanwhile I use –gpus 4 with 4 TITAN V of 12GB MEM free.
I noticed that if I set pretrained_weight to darknet53.hdf5, it can train succeed.
Hint: If you want to see a list of allocated tensors when OOM happens, add report_tensor_allocations_upon_oom to RunOptions for current allocation info.
tensorflow.python.framework.errors_impl.ResourceExhaustedError: OOM when allocating tensor with shape[256,1024,13,13] and type float on /job:localhost/replica:0/task:0/device:GPU:0 by allocator GPU_0_bfc
[[{{node training_1/Adam/gradients/time_distributed_19_1/AvgPool_grad/AvgPoolGrad}}]]
Hint: If you want to see a list of allocated tensors when OOM happens, add report_tensor_allocations_upon_oom to RunOptions for current allocation info.
So I think it may caused by my specs. Here is my specs.
random_seed: 42
enc_key: 'YjlxOTRkaHRjYWI2Z2N***'
verbose: True
network_config {
input_image_config {
image_type: RGB
image_channel_order: 'bgr'
size_height_width {
height: 1080
width: 1920
}
image_channel_mean {
key: 'b'
value: 103.939
}
image_channel_mean {
key: 'g'
value: 116.779
}
image_channel_mean {
key: 'r'
value: 123.68
}
image_scaling_factor: 1.0
max_objects_num_per_image: 100
}
feature_extractor: "darknet:53"
anchor_box_config {
scale: 64.0
scale: 128.0
scale: 256.0
ratio: 1.0
ratio: 0.5
ratio: 2.0
}
freeze_bn: True
freeze_blocks: 0
freeze_blocks: 1
roi_mini_batch: 256
rpn_stride: 16
conv_bn_share_bias: True
roi_pooling_config {
pool_size: 7
pool_size_2x: False
}
all_projections: True
use_pooling:False
}
training_config {
kitti_data_config {
data_sources: {
tfrecords_path: "/workspace/tlt-experiments/tfrecords/kitti_trainval/kitti_trainval*"
image_directory_path: "/workspace/tlt-experiments/data/training"
}
image_extension: 'jpg'
target_class_mapping {
key: 'person'
value: 'person'
}
validation_fold: 0
}
data_augmentation {
preprocessing {
output_image_width: 1920
output_image_height: 1080
output_image_channel: 3
min_bbox_width: 1.0
min_bbox_height: 1.0
}
spatial_augmentation {
hflip_probability: 0.5
vflip_probability: 0.0
zoom_min: 1.0
zoom_max: 1.0
translate_max_x: 0
translate_max_y: 0
}
color_augmentation {
hue_rotation_max: 0.0
saturation_shift_max: 0.0
contrast_scale_max: 0.0
contrast_center: 0.5
}
}
enable_augmentation: True
batch_size_per_gpu: 1
num_epochs: 300
#pretrained_weights: "/workspace/tlt-experiments/data/faster_rcnn/darknet53.hdf5"
resume_from_model: "/workspace/tlt-experiments/data/faster_rcnn/frcnn_kitti_darknet53.epoch111.tlt"
output_model: "/workspace/tlt-experiments/data/faster_rcnn/frcnn_kitti_darknet53.tlt"
rpn_min_overlap: 0.3
rpn_max_overlap: 0.7
classifier_min_overlap: 0.0
classifier_max_overlap: 0.5
gt_as_roi: False
std_scaling: 1.0
classifier_regr_std {
key: 'x'
value: 10.0
}
classifier_regr_std {
key: 'y'
value: 10.0
}
classifier_regr_std {
key: 'w'
value: 5.0
}
classifier_regr_std {
key: 'h'
value: 5.0
}
rpn_mini_batch: 256
rpn_pre_nms_top_N: 12000
rpn_nms_max_boxes: 2000
rpn_nms_overlap_threshold: 0.7
reg_config {
reg_type: 'L2'
weight_decay: 1e-4
}
optimizer {
adam {
lr: 0.00001
beta_1: 0.9
beta_2: 0.999
decay: 0.0
}
}
lr_scheduler {
step {
base_lr: 0.00001
gamma: 1.0
step_size: 30
}
}
lambda_rpn_regr: 1.0
lambda_rpn_class: 1.0
lambda_cls_regr: 1.0
lambda_cls_class: 1.0
inference_config {
#images_dir: '/workspace/tlt-experiments/data/testing/image_2'
images_dir: '/workspace/tlt-experiments/data/lanzhou_test/'
detection_image_output_dir: '/workspace/tlt-experiments/data/faster_rcnn/lanzhou_test_imgs'
labels_dump_dir: '/workspace/tlt-experiments/data/faster_rcnn/lanzhou_test_labels'
model: '/workspace/tlt-experiments/data/faster_rcnn/frcnn_kitti_darknet53.epoch110.tlt'
#detection_image_output_dir: '/workspace/tlt-experiments/data/faster_rcnn/inference_results_imgs'
#labels_dump_dir: '/workspace/tlt-experiments/data/faster_rcnn/inference_dump_labels'
rpn_pre_nms_top_N: 6000
rpn_nms_max_boxes: 300
rpn_nms_overlap_threshold: 0.7
bbox_visualize_threshold: 0.6
classifier_nms_max_boxes: 300
classifier_nms_overlap_threshold: 0.3
}
evaluation_config {
model: '/workspace/tlt-experiments/data/faster_rcnn/frcnn_kitti_darknet53.epoch110.tlt'
labels_dump_dir: '/workspace/tlt-experiments/data/faster_rcnn/test_dump_labels'
rpn_pre_nms_top_N: 6000
rpn_nms_max_boxes: 300
rpn_nms_overlap_threshold: 0.7
classifier_nms_max_boxes: 300
classifier_nms_overlap_threshold: 0.3
object_confidence_thres: 0.0001
use_voc07_11point_metric:False
}
}