From 14581bc12d13cc848cc9167c006454d5c3917e84 Mon Sep 17 00:00:00 2001 From: wangli Date: Sun, 8 Jun 2025 20:25:54 +0800 Subject: [PATCH] update readme --- README.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index fd17bb1..0133fa5 100644 --- a/README.md +++ b/README.md @@ -53,8 +53,6 @@ If you don't have a CUDA environment, you can also try the CPU version. Note tha ## Translation -A language model built with a two-layer decoder, trained on the first 256 tokens from timemachine_preprocessed.txt, reads the text starting from test_lm.txt during inference. - ### training Align the training data volume (512 pairs) of Chapter 11 Transformer in d2l. ``` @@ -142,6 +140,8 @@ i'm home . -> je suis chez moi . ## Language Model +A language model built with a two-layer decoder, trained on the first 256 tokens from timemachine_preprocessed.txt, reads the text starting from test_lm.txt during inference. + ### training ``` $ ./lm -e 10 -m 256