Open mysope opened 7 years ago
This issue is related to https://github.com/tensorflow/tensorflow/issues/7496 . This may fix your errors.
If other errors occur, note that the author @balancap has updated SSD-Tensorflow but not yet pushed changes to this repo. I was experiencing similar errors and updated my SDC-Vehicle-Detection repo by replacing folders nets/
and preprocessing/
with the corresponding versions from SSD-Tensorflow, and adding the directory tf_extended/
to my repo as well. Depending on your use case, you may need to copy a few more files over.
Hi @villanuevab, did you have to modify some code after replacing nets/, preprocessing/ from SSD-Tensorflow repo?
In vehicle-detection.ipynb for example, there are methods that do not exist in the SSD-Tensorflow nets/ such as ssd_common.ssd_bboxes_select. I found that it's equivalent to ssd_common.tf_ssd_bboxes_select but the parameters returned are different.
thanks Peeranat F.
Hello. Thank you for your question! Actually, although I got SSD-Tensorflow to run, I was never able to get good results. Training using this repo (unchanged), SDC-Vehicle-Detection, was more fruitful. I have not kept up to date with this repo or SSD-Tensorflow, but the latest models that actually achieved reasonable results came from using SDC-Vehicle-Detection.
Thanks for your response. So how did you manage to get SDC-Vehicle-Detection to work and get reasonable results and solve "TypeError: init() got multiple values for keyword argument 'dtype'" then?
If you don't mind me asking what model(s) you use that got you the best result for self driving car vehicle detection? My team might try that out. Right now we are focusing on only SSD and Udacity's dataset.
thanks
SSD network construction.
reuse = True if 'ssd' in locals() else None ssd = ssd_vgg_300.SSDNet(params) with slim.arg_scope(ssd.arg_scope(weight_decay=0.0005)): predictions, localisations, logits, end_points = ssd.net(image_4d, is_training=False, reuse=reuse)
When running the above command line on the ipython notebook, the following error occurs. What should I fix?
TypeError Traceback (most recent call last)