This repo provides examples of parameter tuning with DrOpt.
.
|
+- trials/ # parameter tuning trial examples
|
+- data/ # datasets for trials
|
+- util/ # utility Python scirpt
|
+- log/ # trial logs (create upon running)
Our trial examples include the following models:
func-eggholder
: test function for optimizationfunc-rosenbrock
: another test function for optimizationtitanic-xgboost
: Titanic survival prediction by XGBoostmnist-pytorch
: handwritten digit recognition with PyTorch frameworkimagenet-pytorch
: image classification with PyTorch framework
- Set up Python enviroment.
- (Optional) Set up Python virtual enviroment.
- Go to the folder of a trial example.
- Install required Python package:
$ pip install -r requirements.txt
- Run the trial with DrOpt:
$ droptctl -t TOKEN create
- Inspect the trial result on the DrOpt service cloud.
- DrOpt official service cloud: https://dropt.goedge.ai
- DrOpt documentation: https://dropt-cli.readthedocs.io
- dropt-cli source repo: https://github.com/GoEdge-ai/dropt-cli