SSD-Tensorflow

SSD-Tensorflow copied to clipboard

The default value for the 'opt_epsilon' flag is 1.0. Isn't this way too high? The default value for this in tf.train.AdamOptimizer is 1e-8.

anishchopra

Training on new data set

29

Hi, First, thanks for your hard work to implement SSD on TensorFlow, BIG CREDIT! Now I want to train my own model on another data set. 1. It only contains...

Peilin-Yang

hi when in run the commond python train_ssd_network.py --train_dir=${TRAIN_DIR} --dataset_dir=${DATASET_DIR} --dataset_name=pascalvoc_2012 --dataset_split_name=train --model_name=ssd_300_vgg --checkpoint_path=${CHECKPOINT_PATH} --save_summaries_secs=60 --save_interval_secs=600 --weight_decay=0.0005 --optimizer=adam --learning_rate=0.001 --batch_size=32 it shows the error ValueError: Cannot reshape a tensor with...

Davdi

Can I train a num_class is 26(>21) model use the SSD-512 VGG-based pre-trained checkpoints?

19

Can I train a num_class is 36(>21) model use the SSD-512 VGG-based pre-trained checkpoints? when I train a model which num_class is 16(

blacklong28

Finetune using other dataset with NotFoundError: Restoring from checkpoint failed

5

I have met the following problem when I tried to use SSD to detect only the dog class in VOC2007 dataset. So the number of class is two, and I...

xiyuanzh

Per my understanding, the functions bunched together in the sub-directory tf_extended are meant to supplement the SSD implementation using standard TensorFlow functions, but it is **not** the same as [TFX...

mythrandire

Can this model train with 224*224*3 input with good accuracy

parth1595

I0313 15:10:55.695348 140095594268544 learning.py:507] global step 2290: loss = 24.0642 (4.784 sec/step) INFO:tensorflow:Saving checkpoint to path ./train_model/model.ckpt I0313 15:11:12.685011 140091937531648 supervisor.py:1117] Saving checkpoint to path ./train_model/model.ckpt INFO:tensorflow:Error reported to Coordinator:...

haobabuhaoba

I modified the loss function to get good results.

2

Face Detection Data http://vis-www.cs.umass.edu/fddb/  1. Remove the 'crop part' of 'preprocess_for_train()' in ssd_vgg_preprocessing.py While cropping, all true boxes are removed, which causes problems with loss calculations. ``` dst_image, labels,...

hccho2

when I run eval_ssd_network.py,there have wrong? what are in dataset_dir

chenjjjjuu