Add Sentence Transformers usage
#1
by tomaarsen HF Staff - opened
README.md
CHANGED
|
@@ -1,6 +1,7 @@
|
|
| 1 |
---
|
| 2 |
tags:
|
| 3 |
- ColBERT
|
|
|
|
| 4 |
- PyLate
|
| 5 |
- sentence-transformers
|
| 6 |
- sentence-similarity
|
|
@@ -486,14 +487,14 @@ colgrep set-model lightonai/LateOn-Code # default: lightonai/LateOn-Code-edge
|
|
| 486 |
For more information about ColGrep, please refer to the [official documentation](https://github.com/lightonai/next-plaid/tree/main/colgrep)
|
| 487 |
|
| 488 |
|
| 489 |
-
#
|
| 490 |
|
| 491 |
-
This is a
|
| 492 |
|
| 493 |
## Model Details
|
| 494 |
|
| 495 |
### Model Description
|
| 496 |
-
- **Model Type:**
|
| 497 |
<!-- - **Base model:** [Unknown](https://huggingface.co/unknown) -->
|
| 498 |
- **Document Length:** 2048 tokens
|
| 499 |
- **Query Length:** 256 tokens
|
|
@@ -540,6 +541,41 @@ ColBERT(
|
|
| 540 |
```
|
| 541 |
|
| 542 |
## Usage
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 543 |
First install the PyLate library:
|
| 544 |
|
| 545 |
```bash
|
|
|
|
| 1 |
---
|
| 2 |
tags:
|
| 3 |
- ColBERT
|
| 4 |
+
- multi-vector
|
| 5 |
- PyLate
|
| 6 |
- sentence-transformers
|
| 7 |
- sentence-similarity
|
|
|
|
| 487 |
For more information about ColGrep, please refer to the [official documentation](https://github.com/lightonai/next-plaid/tree/main/colgrep)
|
| 488 |
|
| 489 |
|
| 490 |
+
# Model
|
| 491 |
|
| 492 |
+
This is a multi-vector (ColBERT-style late interaction) embedding model finetuned from [lightonai/LateOn-Code-edge-pretrain](https://huggingface.co/lightonai/LateOn-Code-edge-pretrain) on the [apps](https://huggingface.co/datasets/lightonai/nv-embed-supervised-distill-dedup-code), [synthetictext2sql](https://huggingface.co/datasets/lightonai/nv-embed-supervised-distill-dedup-code), [cosqa](https://huggingface.co/datasets/lightonai/nv-embed-supervised-distill-dedup-code), [codefeedbackst](https://huggingface.co/datasets/lightonai/nv-embed-supervised-distill-dedup-code), [codefeedbackmt](https://huggingface.co/datasets/lightonai/nv-embed-supervised-distill-dedup-code), [stackoverflowqa](https://huggingface.co/datasets/lightonai/nv-embed-supervised-distill-dedup-code), [codetranscontest](https://huggingface.co/datasets/lightonai/nv-embed-supervised-distill-dedup-code), [codetransdl](https://huggingface.co/datasets/lightonai/nv-embed-supervised-distill-dedup-code), [CodeSearchNet_go](https://huggingface.co/datasets/lightonai/nv-embed-supervised-distill-dedup-code), [CodeSearchNet_java](https://huggingface.co/datasets/lightonai/nv-embed-supervised-distill-dedup-code), [CodeSearchNet_javascript](https://huggingface.co/datasets/lightonai/nv-embed-supervised-distill-dedup-code), [CodeSearchNet_php](https://huggingface.co/datasets/lightonai/nv-embed-supervised-distill-dedup-code), [CodeSearchNet_python](https://huggingface.co/datasets/lightonai/nv-embed-supervised-distill-dedup-code), [CodeSearchNet_ruby](https://huggingface.co/datasets/lightonai/nv-embed-supervised-distill-dedup-code), [CodeSearchNet_ccr_go](https://huggingface.co/datasets/lightonai/nv-embed-supervised-distill-dedup-code), [CodeSearchNet_ccr_java](https://huggingface.co/datasets/lightonai/nv-embed-supervised-distill-dedup-code), [CodeSearchNet_ccr_javascript](https://huggingface.co/datasets/lightonai/nv-embed-supervised-distill-dedup-code), [CodeSearchNet_ccr_php](https://huggingface.co/datasets/lightonai/nv-embed-supervised-distill-dedup-code), [CodeSearchNet_ccr_python](https://huggingface.co/datasets/lightonai/nv-embed-supervised-distill-dedup-code) and [CodeSearchNet_ccr_ruby](https://huggingface.co/datasets/lightonai/nv-embed-supervised-distill-dedup-code) datasets. It maps sentences & paragraphs to sequences of 48-dimensional dense vectors and can be used for semantic textual similarity using the MaxSim operator.
|
| 493 |
|
| 494 |
## Model Details
|
| 495 |
|
| 496 |
### Model Description
|
| 497 |
+
- **Model Type:** Multi-vector embedding model
|
| 498 |
<!-- - **Base model:** [Unknown](https://huggingface.co/unknown) -->
|
| 499 |
- **Document Length:** 2048 tokens
|
| 500 |
- **Query Length:** 256 tokens
|
|
|
|
| 541 |
```
|
| 542 |
|
| 543 |
## Usage
|
| 544 |
+
|
| 545 |
+
### Sentence Transformers
|
| 546 |
+
|
| 547 |
+
This model can be used with [Sentence Transformers](https://www.sbert.net/) as a multi-vector (ColBERT-style late interaction) retriever via the `MultiVectorEncoder`:
|
| 548 |
+
|
| 549 |
+
```bash
|
| 550 |
+
pip install "sentence-transformers>=6.0.0"
|
| 551 |
+
```
|
| 552 |
+
|
| 553 |
+
```python
|
| 554 |
+
from sentence_transformers import MultiVectorEncoder
|
| 555 |
+
|
| 556 |
+
model = MultiVectorEncoder("lightonai/LateOn-Code-edge")
|
| 557 |
+
|
| 558 |
+
query = "Which planet is known as the Red Planet?"
|
| 559 |
+
documents = [
|
| 560 |
+
"Venus is often called Earth's twin because of its similar size and proximity.",
|
| 561 |
+
"Mars, known for its reddish appearance, is often referred to as the Red Planet.",
|
| 562 |
+
"Jupiter, the largest planet in our solar system, has a prominent red spot.",
|
| 563 |
+
"Saturn, famous for its rings, is sometimes mistaken for the Red Planet.",
|
| 564 |
+
]
|
| 565 |
+
|
| 566 |
+
query_embeddings = model.encode_query(query)
|
| 567 |
+
document_embeddings = model.encode_document(documents)
|
| 568 |
+
print(query_embeddings.shape, document_embeddings[0].shape)
|
| 569 |
+
# (12, 48) (18, 48)
|
| 570 |
+
|
| 571 |
+
# MaxSim late-interaction scoring (higher is more relevant)
|
| 572 |
+
scores = model.similarity(query_embeddings, document_embeddings)
|
| 573 |
+
print(scores)
|
| 574 |
+
# tensor([[4.7185, 7.8521, 7.0537, 7.3523]])
|
| 575 |
+
```
|
| 576 |
+
|
| 577 |
+
### PyLate
|
| 578 |
+
|
| 579 |
First install the PyLate library:
|
| 580 |
|
| 581 |
```bash
|