From f646e846dd85cb15d4f573613f49df69fcfed97f Mon Sep 17 00:00:00 2001 From: Paul Lerner Date: Mon, 16 Dec 2024 15:48:13 +0100 Subject: [PATCH] docs: interact --- README.md | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/README.md b/README.md index 047971b..4bbd681 100644 --- a/README.md +++ b/README.md @@ -259,6 +259,14 @@ filter `A-Za-z` only makes a significant difference for BLOOM which vocabulary i Summing up, you can pass arguments like `python -m neott.viz.pvs bigscience/bloom-7b1 --alpha_filter=true --negatives=all_intra` to only get the relevant metric for BLOOM, and `--alpha_filter=false` for Croissant and Llama. +# Interactive demo + +Use `neott.interact` to interactively translate source terms/generate terms given a definition. + +For example: `python -m neott.interact --config exp/interact.yaml` will apply our ICL method that combines Co-hyponyms and Derivation paradigms, +using FranceTerme as ICL set. You can change the parameters as for `neott.prompt` + + # citation If you use our code or data on Machine Translation, please cite: