From 4c14669a78856dae9bf53e5708935ba397f5e689 Mon Sep 17 00:00:00 2001 From: Patrick von Platen Date: Fri, 18 Sep 2020 16:35:11 +0200 Subject: [PATCH] Update README.md --- model_cards/facebook/rag-token-nq/README.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/model_cards/facebook/rag-token-nq/README.md b/model_cards/facebook/rag-token-nq/README.md index 683414e4f2..0db24fb140 100644 --- a/model_cards/facebook/rag-token-nq/README.md +++ b/model_cards/facebook/rag-token-nq/README.md @@ -1,7 +1,7 @@ ## RAG This is the RAG-Token Model of the the paper [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/pdf/2005.11401.pdf) -by Aleksandra Piktus et al. +by Patrick Lewis, Ethan Perez, Aleksandara Piktus et al. ## Usage: @@ -18,7 +18,7 @@ outputs = model(input_ids=input_dict["input_ids"], labels=input_dict["labels"]) # outputs.loss should give 76.1230 -generated = model.generate(input_ids=input_dict["input_ids"], num_beams=4) +generated = model.generate(input_ids=input_dict["input_ids"]) generated_string = tokenizer.batch_decode(generated, skip_special_tokens=True) # generated_string should give 270,000 -> not quite correct the answer, but it also only uses a dummy index