Object-Detection-API | Yolov3 Object Detection implemented as APIs | Computer Vision library

I'm trying debug my trained Faster R-CNN model using Tensorflow Object Detection API and I want to visualize the proposal regions of RPN on an image. Can anyone tell me how to do it? I found a post here but it hasn't been answered. I tried to export the model using exporter_main_v2.py with only the RPN head as said here and this is the massage when I deleted the second_stage.

I am working on an object detection model in google colab and I'm following most of the instructions outlined here.

In order to train the model, I am trying to use:

While running models/research/object_detection/model_main_tf2.py from tensorflow/models (or just python -c "from object_detection import model_lib_v2") I get:

I followed the "Training Custom Object Detector" tutorial (https://tensorflow-object-detection-api-tutorial.readthedocs.io/en/latest/training.html)

When running the script to continue training a pre-trained model:

python model_main_tf2.py --model_dir=models/my_ssd_resnet50_v1_fpn --pipeline_config_path=models/my_ssd_resnet50_v1_fpn/pipeline.config

(found here: https://github.com/tensorflow/models/blob/master/research/object_detection/model_main_tf2.py)

Using different numpy versions, I get the following errors.

Scenario #1:

• tensorflow: 2.2.0
• numpy: 1.20.0-1

NotImplementedError: Cannot convert a symbolic Tensor (cond_2/strided_slice:0) to a numpy array. This error may indicate that you're trying to pass a Tensor to a NumPy call, which is not supported

I looked online and it suggests to downgrade the numpy version < 1.20.0 (NotImplementedError: Cannot convert a symbolic Tensor (2nd_target:0) to a numpy array). Note, the version numpy version must be >= 1.19.2 for tensorflow 2.2.0.

Scenario #2:

• tensorflow: 2.2.0
• numpy: 1.19.2-5

ValueError: numpy.ndarray size changed, may indicate binary incompatibility. Expected 88 from C header, got 80 from PyObject

However, the online recommendation is to upgrade numpy to >= 1.20.0. (ValueError: numpy.ndarray size changed, may indicate binary incompatibility. Expected 88 from C header, got 80 from PyObject). There is a github issue related to this: https://github.com/tensorflow/models/issues/9749.

I'm not sure what version to use to get the code running.

Appendix: Scenario #1 Error Code (numpy: 1.20.0-1)

I am trying to install Tensor Flow Object Detection on Windows 10.

Resource: https://towardsdatascience.com/how-to-install-tensorflow-2-object-detection-api-on-windows-2eef9b7ae869

After running these steps, we are receiving the errors below. It is stalling on pyarrow. How can this be fixed?

I try to train a model object detection and I follow this tutorial: https://tensorflow-object-detection-api-tutorial.readthedocs.io/en/tensorflow-1.14/training.html

But at the end I execute the command in the cmd : python model_main.py --alsologtostderr --model_dir=training/ --pipeline_config_path=training/ssd_inception_v2_coco.config

and it return the following lines:

I am running confusion matrix on my own custom model using Tensorflow Object Detection API. I am using Faster R-CNN Inception v2 pets. I get this output:

For utilizing Tensorflow's Object Detection transfer learning capabilities, I followed the "Training Custom Object Detector" tutorial (https://tensorflow-object-detection-api-tutorial.readthedocs.io/en/latest/training.html)

When running the script to continue training a pre-trained model:

python model_main_tf2.py --model_dir=models/my_ssd_resnet50_v1_fpn --pipeline_config_path=models/my_ssd_resnet50_v1_fpn/pipeline.config

(found here: https://github.com/tensorflow/models/blob/master/research/object_detection/model_main_tf2.py)

It would give the error failed to create cublas handle: CUBLAS_STATUS_ALLOC_FAILED and Could not create cudnn handle: CUDNN_STATUS_ALLOC_FAILED.

Appendix: Full Error Message.

The accepted answer of this question says how tensorflow draws the bounding boxes of the detected object however does not show or explain how to retrieve these coordinates. Could someone show me how this can be done for tensorflow 2?