No module named 'object_detection'

[Airyzf]

Hi, I got an error when running "python run_stream.py",

Traceback (most recent call last):
File "run_stream.py", line 125, in main
from lib.detection_trt_v1 import TRTV1
File "/home/nvidia/Downloads/realtime_object_detection-master/lib/detection_trt_v1.py", line 4, in
from lib.load_graph_trt_v1 import LoadFrozenGraph
File "/home/nvidia/Downloads/realtime_object_detection-master/lib/load_graph_trt_v1.py", line 10, in
from tf_trt_models.detection import download_detection_model
File "/home/nvidia/Downloads/realtime_object_detection-master/tf_trt_models/detection.py", line 5, in
from object_detection.protos.pipeline_pb2 import TrainEvalPipelineConfig
ImportError: No module named 'object_detection'

I downloaded the object_detection from https://github.com/tensorflow/models and put it to the same directory as the "detection.py",but it didn't work .

[naisy]

Hi @Airyzf,

trt_v1 needs https://github.com/NVIDIA-Jetson/tf_trt_models

And when using PC, you need TensorRT and trt enabled tensorflow (1.8 or later).
Jetson Xavier/TX2, TensorRT is already installed. So you need trt enabled tensorflow (1.8 or later).

[Airyzf]

Hi @Airyzf,

trt_v1 needs https://github.com/NVIDIA-Jetson/tf_trt_models

And when using PC, you need TensorRT and trt enabled tensorflow (1.8 or later).
Jetson Xavier/TX2, TensorRT is already installed. So you need trt enabled tensorflow (1.8 or later).

thank you, it work when the config.yml is nms_v2,
model_type: 'nms_v2'
model_path: 'models/ssd_mobilenet_v1_coco_2018_01_28/frozen_inference_graph.pb'

but dont work trt_v1 , here is my execution step

1. I download https://github.com/NVIDIA-Jetson/tf_trt_models and replace the folder realtime_object_detection/tf_trt_models
2. run the five steps in tf_trt_models Setup
3. download the object_detection from https://github.com/tensorflow/models and put it to realtime_object_detection/tf_trt_models /object_detection
4. edit config.yml
   model_type: 'trt_v1'
   precision_model: 'FP32' # 'FP32', 'FP16', 'INT8'
   model: 'ssdlite_mobilenet_v2_coco_2018_05_09'
5. run "python run_stream.py"

[naisy]

Hi, @Airyzf,

OK, I understood.

1. mkdir ~/github; cd ~/github # base directory. Any directory is OK.
2. git clone https://github.com/naisy/realtime_object_detection
3. git clone https://github.com/NVIDIA-Jetson/tf_trt_models # download to ~/github/tf_trt_models
4. Install tf_trt_models. # When you run install.sh, preparation for https://github.com/tensorflow/models is completed.
5. edit config.yml # Yes, your config.yml
6. run "python run_stream.py" # Yes, your command

ImportError: No module named 'object_detection'

It is installed by tf_trt_models.

[Airyzf]

Hi, @Airyzf,

OK, I understood.

1. mkdir ~/github; cd ~/github # base directory. Any directory is OK.
2. git clone https://github.com/naisy/realtime_object_detection
3. git clone https://github.com/NVIDIA-Jetson/tf_trt_models # download to ~/github/tf_trt_models
4. Install tf_trt_models. # When you run install.sh, preparation for https://github.com/tensorflow/models is completed.
5. edit config.yml # Yes, your config.yml
6. run "python run_stream.py" # Yes, your command

ImportError: No module named 'object_detection'

It is installed by tf_trt_models.

thanks, there is no errors about 'object_detection' , but i got an other error.

Traceback (most recent call last):
File "/home/nvidia/.local/lib/python3.5/site-packages/tensorflow/python/client/session.py", line 1292, in _do_call
return fn(*args)
File "/home/nvidia/.local/lib/python3.5/site-packages/tensorflow/python/client/session.py", line 1277, in _run_fn
options, feed_dict, fetch_list, target_list, run_metadata)
File "/home/nvidia/.local/lib/python3.5/site-packages/tensorflow/python/client/session.py", line 1367, in _call_tf_sessionrun
run_metadata)
tensorflow.python.framework.errors_impl.InvalidArgumentError: Assign requires shapes of both tensors to match. lhs shape= [3,3,576,273] rhs shape= [1,1,576,273]
[[{{node save/Assign_3}} = Assign[T=DT_FLOAT, _class=["loc:@BoxPredictor_0/ClassPredictor/weights"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/device:GPU:0"](BoxPredictor_0/ClassPredictor/weights, save/RestoreV2/_7)]]
[[{{node save/RestoreV2/_16}} = _SendT=DT_FLOAT, client_terminated=false, recv_device="/job:localhost/replica:0/task:0/device:GPU:0", send_device="/job:localhost/replica:0/task:0/device:CPU:0", send_device_incarnation=1, tensor_name="edge_22_save/RestoreV2", _device="/job:localhost/replica:0/task:0/device:CPU:0"]]

[naisy]

Hi @Airyzf,

I cannot check that problem. Because it does not occur in my environment.
The errors seems to related to restore from checkpoint in tf_trt_modes.
tf_trt_models rebuilds the model from the checkpoint. It seems to be an error at that time.

Please check whether tf_trt_models examples can be executed.

[Airyzf]

Hi @Airyzf,

I cannot check that problem. Because it does not occur in my environment.
The errors seems to related to restore from checkpoint in tf_trt_modes.
tf_trt_models rebuilds the model from the checkpoint. It seems to be an error at that time.

Please check whether tf_trt_models examples can be executed.

ok, thank you