trtexec --onnx=resnet50.onnx --shapes=input:1x3x640x640
&&&& RUNNING TensorRT.trtexec # trtexec --onnx=resnet50.onnx --shapes=input:1x3x640x640
[05/13/2021-11:22:59] [I] === Model Options ===
[05/13/2021-11:22:59] [I] Format: ONNX
[05/13/2021-11:22:59] [I] Model: resnet50.onnx
[05/13/2021-11:22:59] [I] Output:
[05/13/2021-11:22:59] [I] === Build Options ===
[05/13/2021-11:22:59] [I] Max batch: explicit
[05/13/2021-11:22:59] [I] Workspace: 16 MiB
[05/13/2021-11:22:59] [I] minTiming: 1
[05/13/2021-11:22:59] [I] avgTiming: 8
[05/13/2021-11:22:59] [I] Precision: FP32
[05/13/2021-11:22:59] [I] Calibration: 
[05/13/2021-11:22:59] [I] Refit: Disabled
[05/13/2021-11:22:59] [I] Safe mode: Disabled
[05/13/2021-11:22:59] [I] Save engine: 
[05/13/2021-11:22:59] [I] Load engine: 
[05/13/2021-11:22:59] [I] Builder Cache: Enabled
[05/13/2021-11:22:59] [I] NVTX verbosity: 0
[05/13/2021-11:22:59] [I] Tactic sources: Using default tactic sources
[05/13/2021-11:22:59] [I] Input(s)s format: fp32:CHW
[05/13/2021-11:22:59] [I] Output(s)s format: fp32:CHW
[05/13/2021-11:22:59] [I] Input build shape: input=1x3x640x640+1x3x640x640+1x3x640x640
[05/13/2021-11:22:59] [I] Input calibration shapes: model
[05/13/2021-11:22:59] [I] === System Options ===
[05/13/2021-11:22:59] [I] Device: 0
[05/13/2021-11:22:59] [I] DLACore: 
[05/13/2021-11:22:59] [I] Plugins:
[05/13/2021-11:22:59] [I] === Inference Options ===
[05/13/2021-11:22:59] [I] Batch: Explicit
[05/13/2021-11:22:59] [I] Input inference shape: input=1x3x640x640
[05/13/2021-11:22:59] [I] Iterations: 10
[05/13/2021-11:22:59] [I] Duration: 3s (+ 200ms warm up)
[05/13/2021-11:22:59] [I] Sleep time: 0ms
[05/13/2021-11:22:59] [I] Streams: 1
[05/13/2021-11:22:59] [I] ExposeDMA: Disabled
[05/13/2021-11:22:59] [I] Data transfers: Enabled
[05/13/2021-11:22:59] [I] Spin-wait: Disabled
[05/13/2021-11:22:59] [I] Multithreading: Disabled
[05/13/2021-11:22:59] [I] CUDA Graph: Disabled
[05/13/2021-11:22:59] [I] Separate profiling: Disabled
[05/13/2021-11:22:59] [I] Skip inference: Disabled
[05/13/2021-11:22:59] [I] Inputs:
[05/13/2021-11:22:59] [I] === Reporting Options ===
[05/13/2021-11:22:59] [I] Verbose: Disabled
[05/13/2021-11:22:59] [I] Averages: 10 inferences
[05/13/2021-11:22:59] [I] Percentile: 99
[05/13/2021-11:22:59] [I] Dump refittable layers:Disabled
[05/13/2021-11:22:59] [I] Dump output: Disabled
[05/13/2021-11:22:59] [I] Profile: Disabled
[05/13/2021-11:22:59] [I] Export timing to JSON file: 
[05/13/2021-11:22:59] [I] Export output to JSON file: 
[05/13/2021-11:22:59] [I] Export profile to JSON file: 
[05/13/2021-11:22:59] [I] 
[05/13/2021-11:22:59] [I] === Device Information ===
[05/13/2021-11:22:59] [I] Selected Device: GeForce RTX 2070 Super
[05/13/2021-11:22:59] [I] Compute Capability: 7.5
[05/13/2021-11:22:59] [I] SMs: 40
[05/13/2021-11:22:59] [I] Compute Clock Rate: 1.38 GHz
[05/13/2021-11:22:59] [I] Device Global Memory: 7973 MiB
[05/13/2021-11:22:59] [I] Shared Memory per SM: 64 KiB
[05/13/2021-11:22:59] [I] Memory Bus Width: 256 bits (ECC disabled)
[05/13/2021-11:22:59] [I] Memory Clock Rate: 7.001 GHz
[05/13/2021-11:22:59] [I] 
----------------------------------------------------------------
Input filename:   resnet50.onnx
ONNX IR version:  0.0.7
Opset version:    13
Producer name:    
Producer version: 
Domain:           
Model version:    0
Doc string:       
----------------------------------------------------------------
[05/13/2021-11:22:59] [W] [TRT] onnx2trt_utils.cpp:220: Your ONNX model has been generated with INT64 weights, while TensorRT does not natively support INT64. Attempting to cast down to INT32.
[05/13/2021-11:22:59] [W] [TRT] onnx2trt_utils.cpp:246: One or more weights outside the range of INT32 was clamped
[05/13/2021-11:22:59] [W] [TRT] onnx2trt_utils.cpp:246: One or more weights outside the range of INT32 was clamped
terminate called after throwing an instance of 'std::out_of_range'
  what():  Attribute not found: axes
Annullato (core dump creato)