Deserialized backend context model.etlt_b16_gpu0_fp16.engine failed to match config params

Accelerated Computing Intelligent Video Analytics DeepStream SDK
, ,
1

I have trained a custom model in nvidia tao and tried to integrate with the deepstream python test2.py, I have edited only the pgie1 and tried to give the custom model path and cfg file. It shows the following error.

I have gone through a similar issue, the answers says check the config file, what is the config file I have to check?


0:00:20.849281607   628      0x1c42610 WARN                 nvinfer gstnvinfer.cpp:643:gst_nvinfer_logger:<secondary1-nvinference-engine> NvDsInferContext[UID 2]: Warning from NvDsInferContextImpl::checkBackendParams() <nvdsinfer_context_impl.cpp:1856> [UID = 2]: backend can not support dims:224x224x3
0:00:20.849307852   628      0x1c42610 ERROR                nvinfer gstnvinfer.cpp:640:gst_nvinfer_logger:<secondary1-nvinference-engine> NvDsInferContext[UID 2]: Error in NvDsInferContextImpl::buildModel() <nvdsinfer_context_impl.cpp:1964> [UID = 2]: deserialized backend context :/opt/nvidia/deepstream/deepstream-6.1/sources/deepstream_python_apps/apps/deepstream-test2-new/deploy/model.etlt_b16_gpu0_fp16.engine failed to match config params
0:00:20.883138286   628      0x1c42610 ERROR                nvinfer gstnvinfer.cpp:640:gst_nvinfer_logger:<secondary1-nvinference-engine> NvDsInferContext[UID 2]: Error in NvDsInferContextImpl::generateBackendContext() <nvdsinfer_context_impl.cpp:2020> [UID = 2]: build backend context failed
0:00:20.883175907   628      0x1c42610 ERROR                nvinfer gstnvinfer.cpp:640:gst_nvinfer_logger:<secondary1-nvinference-engine> NvDsInferContext[UID 2]: Error in NvDsInferContextImpl::initialize() <nvdsinfer_context_impl.cpp:1257> [UID = 2]: generate backend failed, check config file settings
0:00:20.883201356   628      0x1c42610 WARN                 nvinfer gstnvinfer.cpp:846:gst_nvinfer_start:<secondary1-nvinference-engine> error: Failed to create NvDsInferContext instance
0:00:20.883220708   628      0x1c42610 WARN                 nvinfer gstnvinfer.cpp:846:gst_nvinfer_start:<secondary1-nvinference-engine> error: Config file path: dstest2_sgie1_config.txt, NvDsInfer Error: NVDSINFER_CONFIG_FAILED
Error: gst-resource-error-quark: Failed to create NvDsInferContext instance (1): gstnvinfer.cpp(846): gst_nvinfer_start (): /GstPipeline:pipeline0/GstNvInfer:secondary1-nvinference-engine:
Config file path: dstest2_sgie1_config.txt, NvDsInfer Error: NVDSINFER_CONFIG_FAILED

PGIE File

[property]
gpu-id=0
net-scale-factor=1
#model-engine-file=/opt/nvidia/deepstream/deepstream-6.1/sources/deepstream_python_apps/apps/deepstream-test2-new/deploy/model.engine
tlt-encoded-model=/opt/nvidia/deepstream/deepstream-6.1/sources/deepstream_python_apps/apps/deepstream-test2-new/deploy/model.etlt
tlt-model-key=password
uff-input-blob-name=input_1
uff-input-dims=3;224;224;1
int8-calib-file=/opt/nvidia/deepstream/deepstream-6.1/sources/deepstream_python_apps/apps/deepstream-test2-new/deploy/calib.bin
network-input-order=1
#infer-dims=3;224;224
batch-size=16
network-mode=2
num-detected-classes=2
input-object-min-width=10
input-object-min-height=10
model-color-format=1
gpu-id=0
gie-unique-id=2
operate-on-gie-id=1
operate-on-class-ids=0
is-classifier=1
output-blob-names=predictions/Softmax
classifier-async-mode=0
classifier-threshold=0.01
process-mode=2
#scaling-filter=0
#scaling-compute-hw=0

I am sorry to tag you guys, but I need help from you guys.
@yuweiw @yingliu

3

The input dims parameter of your model doesn’t match to the para in your config file. You can check your own models input dim paras.

4

what would be the ideal input dims?
@yuweiw

5

It’s your trained custom model’s input dims.

6

I am using a resnet 18 model with the same dimension. @yuweiw

model_config {
  # Model Architecture can be chosen from:
  # ['resnet', 'vgg', 'googlenet', 'alexnet']
  arch: "resnet"
  # for resnet --> n_layers can be [10, 18, 50]
  # for vgg --> n_layers can be [16, 19]
  n_layers: 18
  use_batch_norm: True
  use_bias: False
  all_projections: False
  use_pooling: True
  retain_head: True
  resize_interpolation_method: BICUBIC
  # if you want to use the pretrained model,
  # image size should be "3,224,224"
  # otherwise, it can be "3, X, Y", where X,Y >= 16
  input_image_size: "3,224,224"
}
8

Close this topic, keep the discussion in Classifier_meta_list is none in deepstream_test2.py - Intelligent Video Analytics / DeepStream SDK - NVIDIA Developer Forums.