Update README.md
Browse files
README.md
CHANGED
|
@@ -113,8 +113,6 @@ High-level training data summary:
|
|
| 113 |
- Data is a mixed corpus pipeline configured in the repository and processed into tokenized shards before training.
|
| 114 |
- SFT stage uses chat/instruction-style datasets with assistant-targeted supervision.
|
| 115 |
|
| 116 |
-
The full training pipeline and dataset composition are described in the repository `README.md`.
|
| 117 |
-
|
| 118 |
All training artifacts are published separately at:
|
| 119 |
|
| 120 |
- [levossadtchi/QED-75M_artifacts](https://huggingface.co/levossadtchi/QED-75M_artifacts)
|
|
|
|
| 113 |
- Data is a mixed corpus pipeline configured in the repository and processed into tokenized shards before training.
|
| 114 |
- SFT stage uses chat/instruction-style datasets with assistant-targeted supervision.
|
| 115 |
|
|
|
|
|
|
|
| 116 |
All training artifacts are published separately at:
|
| 117 |
|
| 118 |
- [levossadtchi/QED-75M_artifacts](https://huggingface.co/levossadtchi/QED-75M_artifacts)
|