Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.
And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.
SSD300 train throw gradients error
See original GitHub issueHi
I tried used the SSD300 model, subsampled the weights to players detection (just 2 labels : persons and background) and train on my own dataset.
When i run model.fit_generator function, i got error 👍 `
An operation has None for gradient. Please make sure that all of your ops have a gradient defined (i.e. are differentiable). Common ops without gradient: K.argmax, K.round, K.eval.
I tried search and solved in the functions definition, but didn’t success with that.
I uploaded my temp code to : https://github.com/amichayfeldman/SSD/blob/master/temp
Any help for that? Thanks
`
Issue Analytics
- State:
- Created 5 years ago
- Comments:7
Top Results From Across the Web
Top Related Medium Post
Top Related StackOverflow Question
Troubleshoot Live Code
Top Related Reddit Thread
Top Related Hackernoon Post
Top Related Tweet
Top Related Dev.to Post
Top Related Hashnode Post
First of all’ try change ‘inference mode’ to ‘training’. Secondly, i iterate over the layers and define which layers i want to train. Somehow this “solved” my issue (i assume that one of the first layers that i finally didnt train caused to the issue)
@amichayfeldman Did you resolve your problem? I have the same problem if I use inference mode and I try to do fit_generator.