Hi Ricardo Andrade, yes, I do plan to complete the series, but it might take a while to publish the blogs. Settling into a new job and a new city takes a lot of your time. Let me know if there is anything, in particular, you need help with, we can discuss it.
I used a GPU instance (p2.xlarge) on AWS with the “deep-learning-for-computer-vision-with-python” AMI. This AMI comes pre-installed with keras-retinanet and other required packages. You can start using the model after activating the RetinaNet virtual environment by
workon retinanet command.
Hi Benthecoder, both are good in their own way as they are built for different purposes. If you read my next post in the series, you can see how I use BeautifulSoup. My general rule is to try to use BeautifulSoup as it is simple and easy to work with. I employ Selenium only when I need to scrape dynamic webpages and emulate human behavior, as I did…
Each row in train.csv (each labeled object) is a training sample for the model. Therefore, steps will be calculated considering the number of rows (total number of objects in all the images), instead of the total number of images.
I don’t think the Generator class method you are talking about is related.
Hi sravan kumar challa,
Look carefully at the lines you added in .bashrc
It should be ‘source /usr/local/bin/virtualenvwrapper.sh’
and not ‘/usr/local/bin/virtualenvwrapper/.sh’
virtualenvwrapper.sh is the file name without the ‘/’ before ‘.sh’