gbure

commit a2553cee69c02fca47496c1cb5c23dc99b5ac5c8
parent 984e8ea40d1b0ce5788af392cfa5b9b67b6f4ea4
Author: Étienne Simon <esimon@esimon.eu>
Date:   Thu, 14 Nov 2019 18:27:06 +0100

README typo base → large

Diffstat:
M
README
 | 
4
++--

1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/README b/README
@@ -7,9 +7,9 @@ $ export DATA_PATH="/tmp" # the directory containing the extracted datasets
 $ export LOG_PATH="/tmp" # the directory where the models will be saved
 $ git clone https://gitlab.lip6.fr/esimon/few-shot-relation-extraction
 $ cd few-shot-relation-extraction
-$ python -m fsre.data.prepare_semeval bert-base-cased
+$ python -m fsre.data.prepare_semeval bert-large-cased
 $ python -m fsre.train fsre/config/soares_supervised_semeval.py
-$ python -m fsre.data.prepare_kbp37 bert-base-cased
+$ python -m fsre.data.prepare_kbp37 bert-large-cased
 $ python -m fsre.train fsre/config/soares_supervised_kbp37.py
 
 I must be missing something since I'm still far from the results reported in the paper: