textcat_architectures

🪐 Weasel Project: Textcat performance benchmarks

Benchmarking different textcat architectures on different datasets.

📋 project.yml

The project.yml defines the data assets required by the project, as well as the available commands and workflows. For details, see the Weasel documentation.

⏯ Commands

The following commands are defined by the project. They can be executed using weasel run [name]. Commands are only re-run if their inputs have changed.

Command	Description
data	Extract the datasets from their archives.
train	Run customized training runs: 3 textcat architectures trained on 2 datasets.
summarize	Summarize the results from the runs and print the best & last scores for each run.

⏭ Workflows

The following workflows are defined by the project. They can be executed using weasel run [name] and will run the specified commands in order. Commands are only re-run if their inputs have changed.

Workflow	Steps
all	data → train → summarize

🗂 Assets

The following assets are defined by the project. They can be fetched by running weasel assets in the project directory.

File	Source	Description
assets/aclImdb_v1.tar.gz	URL	Movie Review Dataset by Maas et al., ACL 2011.
assets/dbpedia_csv.tgz	URL	DBPedia ontology with 14 nonoverlapping classes by Zhang et al., 2015.