SSD-Tensorflow
SSD-Tensorflow copied to clipboard
Single Shot MultiBox Detector in TensorFlow
I try to train my model with train_ssd_network.py and provide true Datasets-DiR, but return an error
tensorflow.python.framework.errors_impl.FailedPreconditionError: tfrecords; Is a directory [[{{node pascalvoc_2007_data_provider/parallel_read/ReaderReadV2_1}}]] This is my .sh: DATASET_DIR=./tfrecords/ TRAIN_DIR=./log/ python3 train_ssd_network.py \ --train_dir=${TRAIN_DIR} \ --dataset_dir=${DATASET_DIR} \ --dataset_name=pascalvoc_2007 \ --dataset_split_name=train \ --model_name=ssd_300_vgg \ --save_summaries_secs=60 \ --save_interval_secs=600 \...
I use this codes to train my own datasets and I have changed some parameters so that it can work on my datasets.However,after running fellowing codes,I meet an error. CHECKPOINT_PATH=./checkpoints/vgg_16.ckpt...
I have trained my own dataset(they are 60 images of 4 classes) as fllow: 1、commd: DATASET_DIR=./tfrecords/voc2007 TRAIN_DIR=./logs/ CHECKPOINT_PATH=./checkpoints/ssd_300_vgg.ckpt python2 train_ssd_network.py \ --train_dir=${TRAIN_DIR} \ --dataset_dir=${DATASET_DIR} \ --dataset_name=pascalvoc_2007 \ --dataset_split_name=train \ --model_name=ssd_300_vgg...
help! which version of tensorflow is ok? what about other packages' version? to train a model by running " python train_ssd_network.py --dataset_dir=**/**/tfrecords "?can't make it
Hello,sir. Can you write the SSD in tensorflow1.13 or newer? When I run eval_ssd_network.py ,it turns out many question of vertion, like 'Use `tf.data.Dataset.from_tensor_slices' or 'TypeError: _variable_v2_call() got an unexpected...
I see '--checkpoint_exclude_scopes=ssd_300_vgg/conv6,ssd_300_vgg/conv7,ssd_300_vgg/block8,ssd_300_vgg/block9,ssd_300_vgg/block10,ssd_300_vgg/block11,ssd_300_vgg/block4_box,ssd_300_vgg/block7_box,ssd_300_vgg/block8_box,ssd_300_vgg/block9_box,ssd_300_vgg/block10_box,ssd_300_vgg/block11_box \ --trainable_scopes=ssd_300_vgg/conv6,ssd_300_vgg/conv7,ssd_300_vgg/block8,ssd_300_vgg/block9,ssd_300_vgg/block10,ssd_300_vgg/block11,ssd_300_vgg/block4_box,ssd_300_vgg/block7_box,ssd_300_vgg/block8_box,ssd_300_vgg/block9_box,ssd_300_vgg/block10_box,ssd_300_vgg/block11_box \', but it doesn't be dealed with in train_ssd_network. Can you give me the complete train_ssd_network.py? Thanks in advance!
**The training parameters are set as follows:** #!/bin/bash DATASET_DIR=../tfrecords/train TRAIN_DIR=../train_model/ CHECKPOINT_PATH=../checkpoints/vgg_16.ckpt python ../train_ssd_network.py \ --train_dir=${TRAIN_DIR} \ --dataset_dir=${DATASET_DIR} \ --dataset_name=pascalvoc_2007 \ --dataset_split_name=train \ --model_name=ssd_300_vgg \ --checkpoint_path=${CHECKPOINT_PATH} \ --checkpoint_model_scope=vgg_16 \ --checkpoint_exclude_scopes=ssd_300_vgg/conv6,ssd_300_vgg/conv7,ssd_300_vgg/block8,ssd_300_vgg/block9,ssd_300_vgg/block10,ssd_300_vgg/block11,ssd_300_vgg/block4_box,ssd_300_vgg/block7_box,ssd_300_vgg/block8_box,ssd_300_vgg/block9_box,ssd_300_vgg/block10_box,ssd_300_vgg/block11_box \...
This code is not reproducible. Why there are so many stars? If you load it from vgg16, the effect of map is poor.
in `/nets/ssd_common.py`, there is a line `mask = tf.logical_and(mask, feat_scores > -0.5)` which is a incorrect method. Look the comments at the front of the code block. > Follow the...