Ideas for tools that would help the end-to-end process for integrating machine learning models into iOS apps
- Swift scripting for data augmentation techniques
- Converting tensor flow model to core ml, integrating all parts of process (inspecting network, etc)
- Setting up tensor flow environment (python virtual env, dependencies & dependency version control, better error output)
- Evaluation tool for if tensor flow layers used will be supported in coreml
- Checking data input/output and inferring conversion parameters (e.g. via code generation)
- Better debugging at earliest stage
- Sanity checking metadata for conversion (e.g. with typing)
- Preparing datasets for use with tensorflow
- Hosting core ml models
- Evaluating performance metrics for core ml models on device
- Resources for iOS for machine learning (Meghan should aggregate the ones she has to start)