Assorted useful tensorflow functions and keras layers
Currently the functionality implemented here includes:
- Basic vector and matrix manipulations
- Matrix exponentiation
- Tools for working with multivariate normal distributions
- Kalman filter-related functionality
- Random rotation matrices
- Tools for sequential data
- Gradient manipulations
- Loss manipulations
- On-line data collection
$ git clone https://github.com/hosford42/TensorFlowUtilities.git
$ cd TensorFlowUtilities/
$ pip3 install .I'm happy to receive pull requests. If you submit new functionality, please ensure your code includes documentation and unit tests. The code you submit will need to be licensed compatibly with the MIT license. If you wrote it yourself, add your name as a contributor. If you didn't, please add the original author as a contributor. If you're looking for ways to contribute, take a look at any open feature requests; these are things I probably won't have time to work on myself.
If you find a bug, I'm eager to hear about it, so please open an issue and report it. I'll try to get to it as soon as I have the time. Unfortunately, for the time being, this is a one-person show, so be prepared to wait -- unless you would like to submit a fix yourself, which would make my day. :)
I'm happy to receive feature requests, but in all honesty, I probably won't get time to implement them unless I have a specific use case for an unrelated project I'm working on.