summaryrefslogtreecommitdiff
path: root/tokenizer
Commit message (Collapse)AuthorAgeFilesLines
* Bump github.com/jonasknobloch/mbpe to v0.1.2Jonas Knobloch2026-06-272-9/+40
|
* Refactor profiling helpers into moduleJonas Knobloch2026-05-131-42/+4
|
* Recover tokenization based on configJonas Knobloch2026-04-244-13/+31
|
* Access artifacts via shelfJonas Knobloch2026-04-241-2/+3
|
* Implement decodingJonas Knobloch2026-04-221-1/+9
|
* Track state in local variableJonas Knobloch2026-04-221-77/+37
|
* Fix FSA staticsJonas Knobloch2026-04-221-1/+1
| | | | * Add missing 've match
* Add sander moduleJonas Knobloch2026-04-171-7/+36
|
* Add token mapping helpersJonas Knobloch2026-04-171-33/+27
|
* Fix reachable merges validatorJonas Knobloch2026-04-141-4/+29
| | | | * Previous implementation ignored byte replacements
* Add lesci moduleJonas Knobloch2026-04-142-0/+34
|
* Return document index in text iteratorJonas Knobloch2026-04-031-5/+1
|
* Scope text column to parquet readerJonas Knobloch2026-04-031-1/+1
|
* Handle unknown runesJonas Knobloch2026-03-313-1/+119
|
* Sort importsJonas Knobloch2026-03-251-0/+1
|
* Tidy dependenciesJonas Knobloch2026-03-252-1/+198
|
* Preserve per-byte error replacementsJonas Knobloch2026-03-244-0/+38
|
* Remove input bufferJonas Knobloch2026-03-242-36/+66
|
* Expose finite state automatonJonas Knobloch2026-03-244-1/+320
|
* Add tokenizer moduleJonas Knobloch2026-03-244-0/+116
* Wrap MBPE tokenizer