Commit
•
57672ab
1
Parent(s):
430ce7c
Update README.md
Browse files
README.md
CHANGED
|
@@ -10,18 +10,38 @@ The prebuilt vocabularies are all available for download [here](https://huggingf
|
|
| 10 |
**July 3:** TokenMonster v1.0 has been released. The "420" prebuilt vocabularies are being released as they are completed, at a rate of around 10 per day. Let me know if there's one you want and I can prioritize it.
|
| 11 |
|
| 12 |
Choose a dataset from:
|
| 13 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 14 |
|
| 15 |
Choose a vocab size from:
|
| 16 |
-
`1024`
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 17 |
|
| 18 |
Choose an optimization mode from:
|
| 19 |
-
`unfiltered`
|
|
|
|
|
|
|
|
|
|
|
|
|
| 20 |
|
| 21 |
For a capcode disabled vocabulary add:
|
| 22 |
-
`nocapcode`
|
| 23 |
|
| 24 |
And finally add the version number:
|
| 25 |
-
`v1`
|
| 26 |
|
| 27 |
-
Examples:
|
|
|
|
|
|
|
|
|
| 10 |
**July 3:** TokenMonster v1.0 has been released. The "420" prebuilt vocabularies are being released as they are completed, at a rate of around 10 per day. Let me know if there's one you want and I can prioritize it.
|
| 11 |
|
| 12 |
Choose a dataset from:
|
| 13 |
+
|
| 14 |
+
- `code`
|
| 15 |
+
- `english`
|
| 16 |
+
- `englishcode`
|
| 17 |
+
- `fiction`
|
| 18 |
|
| 19 |
Choose a vocab size from:
|
| 20 |
+
- `1024`
|
| 21 |
+
- `2048`
|
| 22 |
+
- `4096`
|
| 23 |
+
- `8000`
|
| 24 |
+
- `16000`
|
| 25 |
+
- `24000`
|
| 26 |
+
- `32000`
|
| 27 |
+
- `40000`
|
| 28 |
+
- `50256`
|
| 29 |
+
- `65536`
|
| 30 |
+
- `100256`
|
| 31 |
|
| 32 |
Choose an optimization mode from:
|
| 33 |
+
- `unfiltered`
|
| 34 |
+
- `clean`
|
| 35 |
+
- `balanced`
|
| 36 |
+
- `consistent`
|
| 37 |
+
- `strict`
|
| 38 |
|
| 39 |
For a capcode disabled vocabulary add:
|
| 40 |
+
- `nocapcode`
|
| 41 |
|
| 42 |
And finally add the version number:
|
| 43 |
+
- `v1`
|
| 44 |
|
| 45 |
+
Examples:
|
| 46 |
+
- `fiction-24000-consistent-v1`
|
| 47 |
+
- `code-4096-clean-nocapcode-v1`
|