Discussion group

Discussion group

We'd love to hear from you! Join the discussion group to talk about evaluation strategies for word embeddings, any additional LD scores that should be included in the toolkit, what other extrinsic tasks data should be explored for correlations, and how they may be useful.

In particular, do let us know if you've tried LDT on other models and/or compared with more tasks. Informed choice of best models for task X (and generally, understanding what those black boxes are doing) will become feasible only if there is a large community effort to track what works in what tasks.

If something doesn't work in ldt library, please use the GitHub issue tracker.