Updating the repository to contain the most recent generative model results.. I pushed the curation from 210 to 500 to have a less volatile results when estimating performance on the dev set. Main file to review is called generative_model_benchmarking.ipynb. The 210 extension is just to remind myself that I was using a curated set of 210.
Updating the repository to contain the most recent generative model results.. I pushed the curation from 210 to 500 to have a less volatile results when estimating performance on the dev set. Main file to review is called generative_model_benchmarking.ipynb. The 210 extension is just to remind myself that I was using a curated set of 210.