| 2025-08-24 04:31 | try finetuning encoder-decoder model like t5gemma | vin | 6 | +187 | -26 |
| 2025-08-23 21:23 | add finetuning scripts | vin | 2 | +122 | -0 |
| 2025-08-23 17:35 | add helper script to split dataset.jsonl | vin | 1 | +63 | -0 |
| 2025-08-23 17:29 | add helper script for reviewing input into dataset | vin | 2 | +125 | -1 |
| 2025-08-23 06:26 | adapt asr.py for multiple file inputs | vin | 1 | +30 | -32 |
| 2025-08-23 05:33 | add highlight/visualizer script and update system prompt | vin | 2 | +155 | -30 |
| 2025-08-23 03:28 | whisper.cpp replaced by onnx-asr | vin | 1 | +0 | -42 |
| 2025-08-22 22:36 | update main script | vin | 3 | +134 | -417 |
| 2025-08-22 21:51 | add formatting | vin | 2 | +291 | -3 |
| 2025-08-22 20:53 | improve asr.py | vin | 1 | +65 | -54 |
| 2025-08-22 19:47 | update with newer parakeet | vin | 1 | +1 | -1 |
| 2025-08-13 15:16 | add parakeet asr via onnx | vin | 1 | +68 | -0 |
| 2025-07-19 21:59 | provide better examples in system prompt and retry on error | vin | 1 | +215 | -69 |
| 2025-07-19 16:36 | save output in data/, not where podcasts are | vin | 1 | +14 | -4 |
| 2025-07-19 15:41 | change chunking to overlapping time segments | vin | 1 | +52 | -55 |
| 2025-07-19 07:15 | output data in jsonl | vin | 1 | +6 | -4 |
| 2025-07-19 05:51 | respect ratelimit and change output extension to .gpt | vin | 1 | +38 | -40 |
| 2025-07-19 05:16 | remove prototyping llm prints, use tqdm, use arguments as inputs | vin | 1 | +134 | -140 |
| 2025-07-19 02:45 | initial payload | vin | 2 | +334 | -0 |
