I'm sure I have installed correct version of dependencies ,including Eigen,Sophus,Pangolin and so on.I'm not meeting Errors when "sh make.sh",I also succeed in installing KinectFusion ,that part is OK.But When I download and test the trained model Error occurs.It said "Framebuffer with requested attributes not available. Using available framebuffer. You may see visual artifacts.created window" . The details are as follows:
:~/DA-RNN-master$ ./experiments/scripts/rgbd_scene_multi_rgbd_test.sh 0
- set -e
- export PYTHONUNBUFFERED=True
- PYTHONUNBUFFERED=True
- export CUDA_VISIBLE_DEVICES=0
- CUDA_VISIBLE_DEVICES=0
++ date +%Y-%m-%d_%H-%M-%S
- LOG=experiments/logs/rgbd_scene_multi_rgbd_test.txt.2019-01-11_15-03-36
- exec
++ tee -a experiments/logs/rgbd_scene_multi_rgbd_test.txt.2019-01-11_15-03-36
- echo Logging output to experiments/logs/rgbd_scene_multi_rgbd_test.txt.2019-01-11_15-03-36
Logging output to experiments/logs/rgbd_scene_multi_rgbd_test.txt.2019-01-11_15-03-36
- '[' -f /home/gaochuan/DA-RNN-master/output/rgbd_scene/rgbd_scene_val/vgg16_fcn_rgbd_multi_frame_rgbd_scene_iter_40000/segmentations.pkl ']'
- ./tools/test_net.py --gpu 0 --network vgg16 --model data/fcn_models/rgbd_scene/vgg16_fcn_rgbd_multi_frame_rgbd_scene_iter_40000.ckpt --imdb rgbd_scene_val --cfg experiments/cfgs/rgbd_scene_multi_rgbd.yml --rig data/RGBDScene/camera.json --kfusion 1
registered Linear
registered Linear
registered Poly3
registered Poly3
shapenet_scene_train
shapenet_scene_val
shapenet_single_train
shapenet_single_val
gmu_scene_train
gmu_scene_val
rgbd_scene_train
rgbd_scene_val
rgbd_scene_trainval
lov_train
lov_val
Called with args:
Namespace(cfg_file='experiments/cfgs/rgbd_scene_multi_rgbd.yml', gpu_id=0, imdb_name='rgbd_scene_val', kfusion=True, model='data/fcn_models/rgbd_scene/vgg16_fcn_rgbd_multi_frame_rgbd_scene_iter_40000.ckpt', network_name='vgg16', pretrained_model=None, rig_name='data/RGBDScene/camera.json', wait=True)
Using config:
{'EPS': 1e-14,
'EXP_DIR': 'rgbd_scene',
'FLIP_X': False,
'GPU_ID': 0,
'INPUT': 'RGBD',
'NETWORK': 'VGG16',
'PIXEL_MEANS': array([[[102.9801, 115.9465, 122.7717]]]),
'RNG_SEED': 3,
'ROOT_DIR': '/home/gaochuan/DA-RNN-master',
'TEST': {'GRID_SIZE': 512,
'RANSAC': False,
'SCALES_BASE': [1.0],
'SINGLE_FRAME': False,
'VERTEX_REG': False,
'VISUALIZE': False},
'TRAIN': {'CHROMATIC': True,
'DISPLAY': 20,
'GAMMA': 0.1,
'GRID_SIZE': 512,
'IMS_PER_BATCH': 1,
'LEARNING_RATE': 0.0001,
'MOMENTUM': 0.9,
'NUM_CLASSES': 10,
'NUM_STEPS': 3,
'NUM_UNITS': 64,
'SCALES_BASE': [1.0],
'SINGLE_FRAME': False,
'SNAPSHOT_INFIX': 'rgbd_scene',
'SNAPSHOT_ITERS': 10000,
'SNAPSHOT_PREFIX': 'vgg16_fcn_rgbd_multi_frame',
'STEPSIZE': 30000,
'TRAINABLE': True,
'USE_FLIPPED': False,
'VERTEX_REG': False,
'VERTEX_W': 10.0,
'VISUALIZE': False}}
/gpu:0
Tensor("unstack:0", dtype=float32)
Tensor("unstack_1:0", dtype=float32)
Tensor("conv5_3/conv5_3:0", shape=(?, ?, ?, 512), dtype=float32)
Tensor("conv5_3_p/conv5_3_p:0", shape=(?, ?, ?, 512), dtype=float32)
Tensor("conv4_3/conv4_3:0", shape=(?, ?, ?, 512), dtype=float32)
Tensor("conv4_3_p/conv4_3_p:0", shape=(?, ?, ?, 512), dtype=float32)
Tensor("score_conv4/score_conv4:0", shape=(?, ?, ?, 64), dtype=float32)
Tensor("upscore_conv5_1:0", shape=(?, ?, ?, 64), dtype=float32)
Tensor("fifo_queue_Dequeue:5", dtype=float32)
Tensor("fifo_queue_Dequeue:6", dtype=float32)
Tensor("fifo_queue_Dequeue:7", dtype=float32)
Tensor("unstack_3:0", dtype=float32)
Tensor("unstack_4:0", dtype=float32)
Tensor("upscore_1:0", shape=(?, ?, ?, 64), dtype=float32)
Computeflow(top_data=<tf.Tensor 'flow:0' shape= dtype=float32>, top_weights=<tf.Tensor 'flow:1' shape= dtype=float32>, top_points=<tf.Tensor 'flow:2' shape= dtype=float32>)
Tensor("score/score:0", shape=(?, ?, ?, 10), dtype=float32)
Use network vgg16 in training
2019-01-11 15:03:37.225146: I tensorflow/core/platform/cpu_feature_guard.cc:137] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA
2019-01-11 15:03:37.339921: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:892] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-01-11 15:03:37.340290: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1030] Found device 0 with properties:
name: GeForce RTX 2080 major: 7 minor: 5 memoryClockRate(GHz): 1.71
pciBusID: 0000:01:00.0
totalMemory: 7.76GiB freeMemory: 7.48GiB
2019-01-11 15:03:37.340301: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1120] Creating TensorFlow device (/device:GPU:0) -> (device: 0, name: GeForce RTX 2080, pci bus id: 0000:01:00.0, compute capability: 7.5)
Loading model weights from data/fcn_models/rgbd_scene/vgg16_fcn_rgbd_multi_frame_rgbd_scene_iter_40000.ckpt
rgbd_scene_val
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
{"down":[0,1,0],"forward":[0,0,1],"height":480,"param_names":["fu","fv","u0","v0","k1","k2","k3"],"params":[570.29999999999995,570.29999999999995,320,240,0,0,0],"right":[1,0,0],"serialno":"34178534347","type":"Poly3","width":640}
"Poly3"
params: 570.29999999999995 570.29999999999995 320 240 0 0 0
pose: [
[
1,
0,
0,
0
],
[
0,
1,
0,
0
],
[
0,
0,
1,
0
]
]
{"down":[0,1,0],"forward":[0,0,1],"height":480,"param_names":["fu","fv","u0","v0","k1","k2","k3"],"params":[570.29999999999995,570.29999999999995,320,240,0,0,0],"right":[1,0,0],"serialno":"34178534347","type":"Poly3","width":640}
"Poly3"
params: 570.29999999999995 570.29999999999995 320 240 0 0 0
pose: [
[
1,
0,
0,
0
],
[
0,
1,
0,
0
],
[
0,
0,
1,
0
]
]
T_dc: 1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 1
Framebuffer with requested attributes not available. Using available framebuffer. You may see visual artifacts.created window
./experiments/scripts/rgbd_scene_multi_rgbd_test.sh: 行 25: 15989 段错误 (核心已转储) ./tools/test_net.py --gpu 0 --network vgg16 --model data/fcn_models/rgbd_scene/vgg16_fcn_rgbd_multi_frame_rgbd_scene_iter_40000.ckpt --imdb rgbd_scene_val --cfg experiments/cfgs/rgbd_scene_multi_rgbd.yml --rig data/RGBDScene/camera.json --kfusion 1
"段错误(核心转存储)" means “segmentation fault (core dumped)”. I check the code and I have found in fact Error is at the line 309 of test.py :
if is_kfusion:
KF = kfusion.PyKinectFusion(rig_filename)
But the KinectFusion has been installed correctly so I don't know what causes the Error.
Probable Reason may be:
1.I use RTX 2080 ,with Cuda 8.0 ,Cudnn 6.0.The device is not compatible.
2.I'm using a new computer,I have not installed OpenCV and other libraries in my computer yet.
@yuxng any suggestions? I'm vert interested in this superb project and thanks a lot .
I'm sure I have installed correct version of dependencies ,including Eigen,Sophus,Pangolin and so on.I'm not meeting Errors when "sh make.sh",I also succeed in installing KinectFusion ,that part is OK.But When I download and test the trained model Error occurs.It said "Framebuffer with requested attributes not available. Using available framebuffer. You may see visual artifacts.created window" . The details are as follows:
:~/DA-RNN-master$ ./experiments/scripts/rgbd_scene_multi_rgbd_test.sh 0
++ date +%Y-%m-%d_%H-%M-%S
++ tee -a experiments/logs/rgbd_scene_multi_rgbd_test.txt.2019-01-11_15-03-36
Logging output to experiments/logs/rgbd_scene_multi_rgbd_test.txt.2019-01-11_15-03-36
registered Linear
registered Linear
registered Poly3
registered Poly3
shapenet_scene_train
shapenet_scene_val
shapenet_single_train
shapenet_single_val
gmu_scene_train
gmu_scene_val
rgbd_scene_train
rgbd_scene_val
rgbd_scene_trainval
lov_train
lov_val
Called with args:
Namespace(cfg_file='experiments/cfgs/rgbd_scene_multi_rgbd.yml', gpu_id=0, imdb_name='rgbd_scene_val', kfusion=True, model='data/fcn_models/rgbd_scene/vgg16_fcn_rgbd_multi_frame_rgbd_scene_iter_40000.ckpt', network_name='vgg16', pretrained_model=None, rig_name='data/RGBDScene/camera.json', wait=True)
Using config:
{'EPS': 1e-14,
'EXP_DIR': 'rgbd_scene',
'FLIP_X': False,
'GPU_ID': 0,
'INPUT': 'RGBD',
'NETWORK': 'VGG16',
'PIXEL_MEANS': array([[[102.9801, 115.9465, 122.7717]]]),
'RNG_SEED': 3,
'ROOT_DIR': '/home/gaochuan/DA-RNN-master',
'TEST': {'GRID_SIZE': 512,
'RANSAC': False,
'SCALES_BASE': [1.0],
'SINGLE_FRAME': False,
'VERTEX_REG': False,
'VISUALIZE': False},
'TRAIN': {'CHROMATIC': True,
'DISPLAY': 20,
'GAMMA': 0.1,
'GRID_SIZE': 512,
'IMS_PER_BATCH': 1,
'LEARNING_RATE': 0.0001,
'MOMENTUM': 0.9,
'NUM_CLASSES': 10,
'NUM_STEPS': 3,
'NUM_UNITS': 64,
'SCALES_BASE': [1.0],
'SINGLE_FRAME': False,
'SNAPSHOT_INFIX': 'rgbd_scene',
'SNAPSHOT_ITERS': 10000,
'SNAPSHOT_PREFIX': 'vgg16_fcn_rgbd_multi_frame',
'STEPSIZE': 30000,
'TRAINABLE': True,
'USE_FLIPPED': False,
'VERTEX_REG': False,
'VERTEX_W': 10.0,
'VISUALIZE': False}}
/gpu:0
Tensor("unstack:0", dtype=float32)
Tensor("unstack_1:0", dtype=float32)
Tensor("conv5_3/conv5_3:0", shape=(?, ?, ?, 512), dtype=float32)
Tensor("conv5_3_p/conv5_3_p:0", shape=(?, ?, ?, 512), dtype=float32)
Tensor("conv4_3/conv4_3:0", shape=(?, ?, ?, 512), dtype=float32)
Tensor("conv4_3_p/conv4_3_p:0", shape=(?, ?, ?, 512), dtype=float32)
Tensor("score_conv4/score_conv4:0", shape=(?, ?, ?, 64), dtype=float32)
Tensor("upscore_conv5_1:0", shape=(?, ?, ?, 64), dtype=float32)
Tensor("fifo_queue_Dequeue:5", dtype=float32)
Tensor("fifo_queue_Dequeue:6", dtype=float32)
Tensor("fifo_queue_Dequeue:7", dtype=float32)
Tensor("unstack_3:0", dtype=float32)
Tensor("unstack_4:0", dtype=float32)
Tensor("upscore_1:0", shape=(?, ?, ?, 64), dtype=float32)
Computeflow(top_data=<tf.Tensor 'flow:0' shape= dtype=float32>, top_weights=<tf.Tensor 'flow:1' shape= dtype=float32>, top_points=<tf.Tensor 'flow:2' shape= dtype=float32>)
Tensor("score/score:0", shape=(?, ?, ?, 10), dtype=float32)
Use network
vgg16in training2019-01-11 15:03:37.225146: I tensorflow/core/platform/cpu_feature_guard.cc:137] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA
2019-01-11 15:03:37.339921: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:892] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-01-11 15:03:37.340290: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1030] Found device 0 with properties:
name: GeForce RTX 2080 major: 7 minor: 5 memoryClockRate(GHz): 1.71
pciBusID: 0000:01:00.0
totalMemory: 7.76GiB freeMemory: 7.48GiB
2019-01-11 15:03:37.340301: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1120] Creating TensorFlow device (/device:GPU:0) -> (device: 0, name: GeForce RTX 2080, pci bus id: 0000:01:00.0, compute capability: 7.5)
Loading model weights from data/fcn_models/rgbd_scene/vgg16_fcn_rgbd_multi_frame_rgbd_scene_iter_40000.ckpt
rgbd_scene_val
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
{"down":[0,1,0],"forward":[0,0,1],"height":480,"param_names":["fu","fv","u0","v0","k1","k2","k3"],"params":[570.29999999999995,570.29999999999995,320,240,0,0,0],"right":[1,0,0],"serialno":"34178534347","type":"Poly3","width":640}
"Poly3"
params: 570.29999999999995 570.29999999999995 320 240 0 0 0
pose: [
[
1,
0,
0,
0
],
[
0,
1,
0,
0
],
[
0,
0,
1,
0
]
]
{"down":[0,1,0],"forward":[0,0,1],"height":480,"param_names":["fu","fv","u0","v0","k1","k2","k3"],"params":[570.29999999999995,570.29999999999995,320,240,0,0,0],"right":[1,0,0],"serialno":"34178534347","type":"Poly3","width":640}
"Poly3"
params: 570.29999999999995 570.29999999999995 320 240 0 0 0
pose: [
[
1,
0,
0,
0
],
[
0,
1,
0,
0
],
[
0,
0,
1,
0
]
]
T_dc: 1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 1
Framebuffer with requested attributes not available. Using available framebuffer. You may see visual artifacts.created window
./experiments/scripts/rgbd_scene_multi_rgbd_test.sh: 行 25: 15989 段错误 (核心已转储) ./tools/test_net.py --gpu 0 --network vgg16 --model data/fcn_models/rgbd_scene/vgg16_fcn_rgbd_multi_frame_rgbd_scene_iter_40000.ckpt --imdb rgbd_scene_val --cfg experiments/cfgs/rgbd_scene_multi_rgbd.yml --rig data/RGBDScene/camera.json --kfusion 1
"段错误(核心转存储)" means “segmentation fault (core dumped)”. I check the code and I have found in fact Error is at the line 309 of test.py :
if is_kfusion:
KF = kfusion.PyKinectFusion(rig_filename)
But the KinectFusion has been installed correctly so I don't know what causes the Error.
Probable Reason may be:
1.I use RTX 2080 ,with Cuda 8.0 ,Cudnn 6.0.The device is not compatible.
2.I'm using a new computer,I have not installed OpenCV and other libraries in my computer yet.
@yuxng any suggestions? I'm vert interested in this superb project and thanks a lot .