Notebook showcasing NNI on Surprise SVD #687
Conversation
|
Check out this pull request on ReviewNB: https://app.reviewnb.com/Microsoft/Recommenders/pull/687 Visit www.reviewnb.com to know how we simplify your Jupyter Notebook workflows. |
notebooks/04_model_select_and_optimize/azureml_hyperdrive_surprise_svd.ipynb
Show resolved
Hide resolved
Fixed. |
|
I was trying pytest-httpserver |
|
i'd recommend just leveraging mock to patch requests and provide the desired results, we already have that in our environment. https://stackoverflow.com/questions/15753390/how-can-i-mock-requests-and-the-response |
There was a problem hiding this comment.
apart from some docstrings that are missing, the code looks awesome! great job @anargyri
| "cell_type": "markdown", | ||
| "metadata": {}, | ||
| "source": [ | ||
| "Hyperband follows a different style of configuration from other tuners. See [the NNI documentation](https://nni.readthedocs.io/en/latest/hyperbandAdvisor.html). Note that the [training script](../../reco_utils/nni/svd_training.py) needs to be adjusted as well, since each Hyperband trial receives an additional parameter `STEPS`, which corresponds to the resource allocation _r<sub>i</sub>_ in the [Hyperband algorithm](https://arxiv.org/pdf/1603.06560.pdf). In this example, we used `STEPS` in combination with `R` to determine the number of epochs that SVD will run for in every trial." |
There was a problem hiding this comment.
We will change STEPS to TRIAL_BUDGET for easy understanding in the next release v0.7, very sorry for this broken change. We will inform you when v0.7 is released. Planed to be in next week (microsoft/nni#963). Thanks.
There was a problem hiding this comment.
Thanks for the notice
|
I added tests except |
Notebook showcasing NNI on Surprise SVD
Description
NNI is a toolkit for hyperparameter tuning that incorporates several tuning methods from the literature.
This notebook shows how to apply NNI for tuning Surprise SVD on Movielens100K.
Checklist: