Update README.md
Browse files
README.md
CHANGED
|
@@ -16,6 +16,7 @@ from the VDocRetriever Phi-3-Vision adapter. This repository contains the traine
|
|
| 16 |
and ConceptFormer sidecar, not a duplicate of the approximately 7.8 GB base model. Loaders
|
| 17 |
resolve the base model from the `base_model` metadata and merge this adapter at load time.
|
| 18 |
|
|
|
|
| 19 |
## Configuration
|
| 20 |
|
| 21 |
- Latent concept token: `<|lcon|>`
|
|
|
|
| 16 |
and ConceptFormer sidecar, not a duplicate of the approximately 7.8 GB base model. Loaders
|
| 17 |
resolve the base model from the `base_model` metadata and merge this adapter at load time.
|
| 18 |
|
| 19 |
+
Paper can be seen in [Arxiv](https://arxiv.org/abs/2608.15698)
|
| 20 |
## Configuration
|
| 21 |
|
| 22 |
- Latent concept token: `<|lcon|>`
|