Text Retrieval
Safetensors
sentence-transformers
English
PyLate
modernbert
ColBERT
sentence-similarity
feature-extraction
late-interaction
reasoning-retrieval
edge
loss:CachedContrastive
Instructions to use DataScience-UIBK/SmallReason-ColBERT-32M with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- sentence-transformers
How to use DataScience-UIBK/SmallReason-ColBERT-32M with sentence-transformers:
from pylate import models queries = [ "Which planet is known as the Red Planet?", "What is the largest planet in our solar system?", ] documents = [ ["Mars is the Red Planet.", "Venus is Earth's twin."], ["Jupiter is the largest planet.", "Saturn has rings."], ] model = models.ColBERT(model_name_or_path="DataScience-UIBK/SmallReason-ColBERT-32M") queries_emb = model.encode(queries, is_query=True) docs_emb = model.encode(documents, is_query=False) - Notebooks
- Google Colab
- Kaggle
Clarify head-loading behaviour
Browse files
README.md
CHANGED
|
@@ -41,8 +41,10 @@ sentence-transformers load **silently ignores it** and gives you the un-headed b
|
|
| 41 |
| `pylate.models.ColBERT(...)` — plain load | base only, head ignored, **no error** | 19.61 |
|
| 42 |
| `WeightedColBERT.from_base(...)` — see below | full model | **21.41** |
|
| 43 |
|
| 44 |
-
There is no warning when the head is skipped
|
| 45 |
-
use the second path.
|
|
|
|
|
|
|
| 46 |
|
| 47 |
---
|
| 48 |
|
|
|
|
| 41 |
| `pylate.models.ColBERT(...)` — plain load | base only, head ignored, **no error** | 19.61 |
|
| 42 |
| `WeightedColBERT.from_base(...)` — see below | full model | **21.41** |
|
| 43 |
|
| 44 |
+
There is no warning when the head is skipped, so if you are reproducing the paper
|
| 45 |
+
number, use the second path. `WeightedColBERT.from_base` resolves the head from this
|
| 46 |
+
repo automatically and **raises** if it cannot find one, so that path cannot fail
|
| 47 |
+
silently. Pass `require_head=False` if you deliberately want the base.
|
| 48 |
|
| 49 |
---
|
| 50 |
|