haydn-jones
commited on
Commit
•
d647bb8
1
Parent(s):
ccb0f1c
Update README.md
Browse files
README.md
CHANGED
@@ -43,4 +43,4 @@ Every tranch was downloaded, this is not the ~1B example ML subset from https://
|
|
43 |
|
44 |
The dataset was entirely shuffled then split into 80%/10%/10% splits for train/val/test.
|
45 |
|
46 |
-
A file vocab.
|
|
|
43 |
|
44 |
The dataset was entirely shuffled then split into 80%/10%/10% splits for train/val/test.
|
45 |
|
46 |
+
A file vocab.csv is in the root of the reposity that contains all of the SELFIES tokens found in the data, with [START], [STOP], and [PAD] added.
|