Please provide the following info (tick the boxes after creating this topic):
Software Version
DRIVE OS 6.0.6
DRIVE OS 6.0.5
DRIVE OS 6.0.4 (rev. 1)
[*] DRIVE OS 6.0.4 SDK
other
Target Operating System
[*] Linux
QNX
other
Hardware Platform
DRIVE AGX Orin Developer Kit (940-63710-0010-D00)
DRIVE AGX Orin Developer Kit (940-63710-0010-C00)
[*] DRIVE AGX Orin Developer Kit (not sure its number)
other
SDK Manager Version
1.9.2.10884
[*] other
Host Machine Version
native Ubuntu Linux 20.04 Host installed with SDK Manager
native Ubuntu Linux 20.04 Host installed with DRIVE OS Docker Containers
native Ubuntu Linux 18.04 Host installed with DRIVE OS Docker Containers
[*] other
I ran a command
./cuDLAStandaloneMode …/model_test/joint_model_cygnus_simple_batch1_512960_230220_dla_opt_sub_1.cudla 2
The output tensor information is as follows
Printing input tensor descriptor
TENSOR NAME : input’
size: 1966080
dims: [1, 4, 512, 960]
data fmt: 2
data type: 4
data category: 0
pixel fmt: 12
pixel mapping: 0
stride[0]: 1
stride[1]: 3840
stride[2]: 0
stride[3]: 0
-----------------------------------------tensor input:0
Printing output tensor descriptor
TENSOR NAME : M0_519’
size: 131072
dims: [1, 64, 32, 60]
data fmt: 1
data type: 4
data category: 3
pixel fmt: 0
pixel mapping: 0
stride[0]: 1
stride[1]: 64
stride[2]: 2048
stride[3]: 2048
-----------------------------------------tensor output:0
CUDLA_DATA_FORMAT_NHWC,The format of the input tensor is CUDLA_DATA_FORMAT_NHWC and the type is int8.
I have an input to an NCHW float, and I should get the quantization parameters and write them into input tensor in CUDLA_DATA_FORMAT_NHWC format.I don’t quite understand what this street means.
Then where can I get the quantized parameters of Tensor.
The output tensort is int8, CUDLA_ DATA_ FORMAT_ NCHW format. How should I inversely convert it to float 32 and convert it to nchw format.
I didn’t find its scale and zero in the model for either Tensor_ Parameter of point