Please provide the following information when requesting support.
• Hardware (T4/V100/Xavier/Nano/etc)
Jetson Xavier
• Network Type (Detectnet_v2/Faster_rcnn/Yolo_v4/LPRnet/Mask_rcnn/Classification/etc)
GazeNet
• TLT Version (Please run “tlt info --verbose” and share “docker_tag” here)
8.2.1
• Training spec file(If have, please share here)
• How to reproduce the issue ? (This is for errors. Please share the command line and the detailed log here.)
Using the deepstream_tao_apps models and tao_converter, I’m trying to generate the model engine file manually.
Start command:
sudo ./tao-converter -e /opt/nvidia/deepstream/deepstream/sources/deepstream_tao_apps/models/gazenet/gazenet_facegrid.etlt_b8_gpu0_fp16.engine \
-p input_right_images:0,1x1x224x224,8x1x224x224,8x1x224x224 \
-p input_facegrid:0,1x1x625x1,8x1x625x1,8x1x625x1 \
-p input_face_images:0,1x1x224x224,8x1x224x224,8x1x224x224 \
-p input_left_images:0,1x1x224x224,8x1x224x224,8x1x224x224 \
-t fp16 -k nvidia_tlt -m 8 nvidia_tlt /opt/nvidia/deepstream/deepstream/sources/deepstream_tao_apps/models/gazenet/gazenet_facegrid.etlt
output:
[INFO] [MemUsageChange] Init CUDA: CPU +363, GPU +0, now: CPU 381, GPU 14483 (MiB)
[INFO] [MemUsageSnapshot] Begin constructing builder kernel library: CPU 381 MiB, GPU 14484 MiB
[INFO] [MemUsageSnapshot] End constructing builder kernel library: CPU 486 MiB, GPU 14605 MiB
[INFO] ----------------------------------------------------------------
[INFO] Input filename: /tmp/filetyyxGm
[INFO] ONNX IR version: 0.0.5
[INFO] Opset version: 10
[INFO] Producer name: tf2onnx
[INFO] Producer version: 1.6.3
[INFO] Domain:
[INFO] Model version: 0
[INFO] Doc string:
[INFO] ----------------------------------------------------------------
[WARNING] onnx2trt_utils.cpp:366: Your ONNX model has been generated with INT64 weights, while TensorRT does not natively support INT64. Attempting to cast down to INT32.
[WARNING] ShapedWeights.cpp:173: Weights fg_fc1_dense/kernel:0 has been transposed with permutation of (1, 0)! If you plan on overwriting the weights with the Refitter API, the new weights must be pre-transposed.
[WARNING] ShapedWeights.cpp:173: Weights fg_fc2_dense/kernel:0 has been transposed with permutation of (1, 0)! If you plan on overwriting the weights with the Refitter API, the new weights must be pre-transposed.
[WARNING] ShapedWeights.cpp:173: Weights fc1_e1_dense/kernel:0 has been transposed with permutation of (1, 0)! If you plan on overwriting the weights with the Refitter API, the new weights must be pre-transposed.
[WARNING] ShapedWeights.cpp:173: Weights fc1_f_dense/kernel:0 has been transposed with permutation of (1, 0)! If you plan on overwriting the weights with the Refitter API, the new weights must be pre-transposed.
[WARNING] ShapedWeights.cpp:173: Weights fc2_f_dense/kernel:0 has been transposed with permutation of (1, 0)! If you plan on overwriting the weights with the Refitter API, the new weights must be pre-transposed.
[WARNING] ShapedWeights.cpp:173: Weights xyz_fc_top_dense/kernel:0 has been transposed with permutation of (1, 0)! If you plan on overwriting the weights with the Refitter API, the new weights must be pre-transposed.
[WARNING] ShapedWeights.cpp:173: Weights xyz_fc/kernel:0 has been transposed with permutation of (1, 0)! If you plan on overwriting the weights with the Refitter API, the new weights must be pre-transposed.
[WARNING] ShapedWeights.cpp:173: Weights tp_fc_top_dense/kernel:0 has been transposed with permutation of (1, 0)! If you plan on overwriting the weights with the Refitter API, the new weights must be pre-transposed.
[WARNING] ShapedWeights.cpp:173: Weights tp_fc/kernel:0 has been transposed with permutation of (1, 0)! If you plan on overwriting the weights with the Refitter API, the new weights must be pre-transposed.
[INFO] Detected input dimensions from the model: (-1, 1, 224, 224)
[INFO] Detected input dimensions from the model: (-1, 1, 625, 1)
[INFO] Detected input dimensions from the model: (-1, 1, 224, 224)
[INFO] Detected input dimensions from the model: (-1, 1, 224, 224)
[INFO] Model has dynamic shape. Setting up optimization profiles.
[INFO] Using optimization profile min shape: (1, 1, 224, 224) for input: input_right_images:0
[INFO] Using optimization profile opt shape: (8, 1, 224, 224) for input: input_right_images:0
[INFO] Using optimization profile max shape: (8, 1, 224, 224) for input: input_right_images:0
[INFO] Using optimization profile min shape: (1, 625, 1) for input: input_facegrid:0
[INFO] Using optimization profile opt shape: (1, 625, 1) for input: input_facegrid:0
[INFO] Using optimization profile max shape: (1, 625, 1) for input: input_facegrid:0
[INFO] Using optimization profile min shape: (1, 224, 224) for input: input_face_images:0
[INFO] Using optimization profile opt shape: (1, 224, 224) for input: input_face_images:0
[INFO] Using optimization profile max shape: (1, 224, 224) for input: input_face_images:0
[INFO] Using optimization profile min shape: (1, 224, 224) for input: input_left_images:0
[INFO] Using optimization profile opt shape: (1, 224, 224) for input: input_left_images:0
[INFO] Using optimization profile max shape: (1, 224, 224) for input: input_left_images:0
[WARNING] DLA requests all profiles have same min, max, and opt value. All dla layers are falling back to GPU
[ERROR] 4: [network.cpp::validate::2951] Error Code 4: Internal Error (input_left_images:0: number of dimensions is 4 but profile 0 has 3.)
[ERROR] Unable to create engine
Segmentation fault
input file:
gazenet_facegrid.etlt (17.3 MB)
