New experiment run command.
#37
Draft
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This commit adds new `experiment run` command that simplifies (to some extent) the process of running ML experiments. This command accepts only one argument - path to the experiment configuration file in YAML format. Previous mechanism to run experiments (experiment `experiment train` and `experiment search_hp`) is still supported, but will likely be removed in future releases. For more details, run the following command: ```shell python -m xtime.main experiment --help ``` 1. Create a template experiment configuration file by running ```shell python -m xtime.main experiment create EXPERIMENT_TYPE EXPERIMENT_DEFINITION_FILE ``` where EXPERIMENT_TYPE is the name of one of supported experiments (`train` or `search_hp`) and EXPERIMENT_DEFINITION_FILE is the path to the YAML file that will be created. 2. Edit that file. The source code of these two experiments provides detailed description of each parameters in this file, but many of them have the same semantics as current CLI API. 3. Run experiment by executing the following command: ```shell python -m xtime.main EXPERIMENT_DEFINITION_FILE ``` This file be serialized in experiment's artifact directory, so it can later be reproduced. This commit also fixes several bugs: - In some cases, the code ignores the experiment name provided by users and uses the `default` experiment. This is fixed now. This commit adds new dependencies: - New python dependency - `OmegaConf` library to interact with configurations of experiments.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Related Issues / Pull Requests
NA
Description
This commit adds new
experiment runcommand that simplifies (to some extent) the process of running ML experiments. This command accepts only one argument - path to the experiment configuration file in YAML format. Previous mechanism to run experiments (experiment trainandexperiment search_hp) is still supported, but will likely be removed in future releases. For more details, run the following command:trainorsearch_hp) and EXPERIMENT_DEFINITION_FILE is the path to the YAML file that will be created.This file be serialized in experiment's artifact directory, so it can later be reproduced.
This commit also fixes several bugs:
defaultexperiment. This is fixed now.This commit adds new dependencies / features:
OmegaConflibrary to interact with configurations of experiments.What changes are proposed in this pull request?
Checklist: