object-detection-api

问题 I have trained a faster rcnn model with a custom dataset using Tensorflow's Object Detection Api. Over time I would like to continue to update the model with additional images (collected weekly). The goal is to optimize for accuracy and to weight newer images over time. Here are a few alternatives: Add images to previous dataset and train a completely new model Add images to previous dataset and continue training previous model New dataset with just new images and continue training previous

问题 I have trained a faster rcnn model with a custom dataset using Tensorflow's Object Detection Api. Over time I would like to continue to update the model with additional images (collected weekly). The goal is to optimize for accuracy and to weight newer images over time. Here are a few alternatives: Add images to previous dataset and train a completely new model Add images to previous dataset and continue training previous model New dataset with just new images and continue training previous

问题 I have followed this tutorial to retrain MobileNet SSD V1 using Tensorflow GPU as described and got 0.5 loss after training using GPU (below more info about config) and got model.ckpt . This is the command I used for Training: python ../models/research/object_detection/legacy/train.py --logtostderr --train_dir=./data/ --pipeline_config_path=./ssd_mobilenet_v1_pets.config And this is the command for freezing (generate pb file): python ../models/research/object_detection/export_inference_graph

问题 I'm fine-tuning SSD object detector using TensorFlow object detection API on Open Images Dataset. My training data contains imbalanced classes, e.g. top (5K images) dress (50K images) etc... I would like to add class weights to classification loss to improve performance. How do I do that? The following section of the config file seems relevant: loss { classification_loss { weighted_sigmoid { } } localization_loss { weighted_smooth_l1 { } } ... classification_weight: 1.0 localization_weight: 1

问题 I am using Tensorflow's Object Detection API in training. In which file, the freezed layers are defined to fine-tune the model in training. I need to experiment changing freezed layers in fine-tuning. For example, if I use Resnet50 configuration, where I can change the freezed layers? 回答1: That certainly you can do. By reading the proto file for training, there is a field called freeze_variables , this is supposed to be a list containing all variables that you want to freeze, e.g. excluding

问题 I am currently working with the Tensorflow Object-Detection API and I want to fine-tune a pre-trained model. Therefore, a hyperparameter-tuning is required. Does the API already provide some kind of hyperparameter-tuning (like a grid search)? If there is nothing available, how can I implement a simple grid search to tune (the most relevant) hyperparameters? Furthermore, does the API provide some kind of Early Stopping function that automatically aborts the training process if the accuracy

问题 I have followed the entire steps/format of codes(cross checked multiple times to be 100% sure they are correct) and the required data for training custom objects on Tensorflow Object Detection API. I tried using ssd_mobilenet_v1_coco, faster_rcnn_resnet101_coco as well as faster_rcnn_inception_v2_coco models and still haven't got any sort of good result. All I get is misclassification of objects or no bounding box at all. I am training to detect a single class object with number of training

问题 I am implementing a Faster RCNN v2 Inception in Tensorflow Object Detection API. To remove redundant overlapping detections, I read that NMS should be applied. One way of doing this is adjusting the NMS IOU Threshold in the config file first_stage_nms_iou_threshold . Questions What is this parameter exactly? To what value should this parameter be adjusted to (default value is 0.7) Why is it called first_stage_nms_iou_threshold ? Why first stage only? Is there another easy and more effective

问题 In their respective config files, by default the faster-RCNN s have only random horizontal flips enabled and the SSD s have random horizontal flips followed by SSD random crop enabled. I want to add more augmentation options. I wrote the following snippet to do so. import tensorflow as tf from object_detection.protos import pipeline_pb2, preprocessor_pb2 from google.protobuf import text_format def get_configs_from_pipeline_file(pipeline_config_path): pipeline_config = pipeline_pb2

问题 I will train my dataset with faster-rcnn for one class. All my images are 1920x1080 sizes. Should I resize or crop the images or I can train with this size? Also my objects are really small (around 60x60). In the config file there are dimensions written as min_dimension: 600 and max_dimension: 1024 for this reason I am confused to train the model with 1920x1080 size images. 回答1: If your objects are small, resizing the images to a smaller size is not a good idea. You can change the max