Spaces:
Running
Running
Antoine Chaffin
commited on
Commit
·
527fc59
1
Parent(s):
d3e89bb
Final touch
Browse files
app.py
CHANGED
|
@@ -37,10 +37,8 @@ readme_md = """
|
|
| 37 |
# MCPyLate
|
| 38 |
|
| 39 |
Multi-vector search has shown very strong performance compared to single dense vector search in numerous domain, including out-of-domain, [long-context](https://x.com/antoine_chaffin/status/1919396926736257521) and [reasoning-intensive](https://x.com/antoine_chaffin/status/1925555110521798925) retrieval.
|
| 40 |
-
|
| 41 |
They are thus particularly well suited for modern retrieval use cases, including agentic workflows.
|
| 42 |
[PyLate](https://lightonai.github.io/pylate/) is library built on top of sentence-transformers that allows to easily train and use multi-vector models.
|
| 43 |
-
|
| 44 |
This MCP server is a demonstration of the use of PyLate models alongside its index optimized for multi-vector search, PLAID.
|
| 45 |
|
| 46 |
This repository propose to search among the [leetcode split of the BRIGHT dataset](https://huggingface.co/datasets/xlangai/BRIGHT/viewer/documents/leetcode) using [Reason-ModernColBERT](https://huggingface.co/lightonai/Reason-ModernColBERT). This 150M parameters model outperforms 7B models on this reasoning-intensive retrieval benchmark, which requires reasoning to find relevant documents.
|
|
|
|
| 37 |
# MCPyLate
|
| 38 |
|
| 39 |
Multi-vector search has shown very strong performance compared to single dense vector search in numerous domain, including out-of-domain, [long-context](https://x.com/antoine_chaffin/status/1919396926736257521) and [reasoning-intensive](https://x.com/antoine_chaffin/status/1925555110521798925) retrieval.
|
|
|
|
| 40 |
They are thus particularly well suited for modern retrieval use cases, including agentic workflows.
|
| 41 |
[PyLate](https://lightonai.github.io/pylate/) is library built on top of sentence-transformers that allows to easily train and use multi-vector models.
|
|
|
|
| 42 |
This MCP server is a demonstration of the use of PyLate models alongside its index optimized for multi-vector search, PLAID.
|
| 43 |
|
| 44 |
This repository propose to search among the [leetcode split of the BRIGHT dataset](https://huggingface.co/datasets/xlangai/BRIGHT/viewer/documents/leetcode) using [Reason-ModernColBERT](https://huggingface.co/lightonai/Reason-ModernColBERT). This 150M parameters model outperforms 7B models on this reasoning-intensive retrieval benchmark, which requires reasoning to find relevant documents.
|