| Commit message (Collapse) | Author | Age | Files | Lines | |
|---|---|---|---|---|---|
| * | Return document index in text iterator | Jonas Knobloch | 2026-04-03 | 2 | -12/+13 |
| | | |||||
| * | Scope text column to parquet reader | Jonas Knobloch | 2026-04-03 | 2 | -2/+2 |
| | | |||||
| * | Use token buffer to generate jobs | Jonas Knobloch | 2026-04-03 | 2 | -111/+36 |
| | | |||||
| * | Return relative offset in push iterator | Jonas Knobloch | 2026-04-03 | 2 | -28/+32 |
| | | |||||
| * | Refactor tail cleanup | Jonas Knobloch | 2026-04-03 | 1 | -13/+13 |
| | | |||||
| * | Track already seen tokens | Jonas Knobloch | 2026-04-02 | 2 | -9/+75 |
| | | |||||
| * | Add token buffer | Jonas Knobloch | 2026-04-02 | 2 | -0/+250 |
| | | |||||
| * | Count skipped texts | Jonas Knobloch | 2026-03-31 | 1 | -2/+14 |
| | | |||||
| * | Refactor llm module | Jonas Knobloch | 2026-03-25 | 2 | -0/+0 |
| | | |||||
| * | Expose progress bar title | Jonas Knobloch | 2026-03-25 | 4 | -6/+6 |
| | | |||||
| * | Add helper to collect results | Jonas Knobloch | 2026-03-25 | 3 | -44/+39 |
| | | |||||
| * | Refactor evaluator initialization | Jonas Knobloch | 2026-03-25 | 4 | -27/+9 |
| | | |||||
| * | Sort imports | Jonas Knobloch | 2026-03-25 | 1 | -2/+1 |
| | | |||||
| * | Tidy dependencies | Jonas Knobloch | 2026-03-25 | 2 | -1/+17 |
| | | |||||
| * | Refactor tokenizer usage | Jonas Knobloch | 2026-03-24 | 1 | -12/+7 |
| | | |||||
| * | Refactor progress bar life cycle | Jonas Knobloch | 2026-03-24 | 1 | -6/+3 |
| | | |||||
| * | Sync dependencies | Jonas Knobloch | 2026-03-23 | 2 | -44/+26 |
| | | |||||
| * | Add log probs extractor | Jonas Knobloch | 2026-03-17 | 4 | -35/+355 |
| | | |||||
| * | Update perplexity evaluator | Jonas Knobloch | 2026-03-17 | 9 | -208/+304 |
| | | | | | | | * Introduce results channel * Introduce jobs channel * Generalize evaluation via callback | ||||
| * | Rename dataset reader | Jonas Knobloch | 2026-03-16 | 2 | -4/+4 |
| | | |||||
| * | Use vanity import path | Jonas Knobloch | 2026-03-16 | 3 | -5/+7 |
| | | |||||
| * | Add ppl Python script | Jonas Knobloch | 2026-03-09 | 1 | -0/+112 |
| | | |||||
| * | Fix perplexity calculation | Jonas Knobloch | 2026-03-09 | 4 | -12/+31 |
| | | | | | | * Track total evaluated tokens * Use first token as context | ||||
| * | Expose window and stride | Jonas Knobloch | 2026-03-09 | 2 | -7/+7 |
| | | |||||
| * | Add ppl command | Jonas Knobloch | 2026-03-09 | 1 | -0/+69 |
| | | |||||
| * | Use dataset package | Jonas Knobloch | 2026-03-09 | 1 | -16/+13 |
| | | |||||
| * | Update module paths | Jonas Knobloch | 2026-02-28 | 1 | -1/+1 |
| | | |||||
| * | Bump Go to 1.25.0 | Jonas Knobloch | 2026-02-10 | 1 | -1/+1 |
| | | |||||
| * | Bump github.com/jonasknobloch/mbpe to v0.1.1 | Jonas Knobloch | 2026-02-10 | 3 | -1/+89 |
| | | |||||
| * | Wait for progress bar | Jonas Knobloch | 2025-11-26 | 1 | -0/+2 |
| | | |||||
| * | Use seen tokens to warm up cache | Jonas Knobloch | 2025-11-26 | 1 | -2/+2 |
| | | |||||
| * | Add perplexity evaluation | Jonas Knobloch | 2025-11-18 | 5 | -0/+340 |
| | | |||||
| * | Scaffold byte-pair correction | Jonas Knobloch | 2025-11-17 | 3 | -0/+13 |
