sergioburdisso
commited on
Commit
•
ea8ba18
1
Parent(s):
174051a
Update README.md
Browse files
README.md
CHANGED
@@ -28,8 +28,9 @@ widget:
|
|
28 |
score: 0.27
|
29 |
---
|
30 |
|
|
|
31 |
|
32 |
-
# Dialog2Flow single target (DSE-base)
|
33 |
|
34 |
This a variation of the **D2F$_{single}$** model introduced in the paper ["Dialog2Flow: Pre-training Soft-Contrastive Action-Driven Sentence Embeddings for Automatic Dialog Flow Extraction"](https://arxiv.org/abs/2410.18481) published in the EMNLP 2024 main conference.
|
35 |
This version uses DSE-base as the backbone model which yields to an increase in performance as compared to the vanilla version using BERT-base as the backbone (results reported in Appendix C).
|
|
|
28 |
score: 0.27
|
29 |
---
|
30 |
|
31 |
+
![image/png](voronoi_umap.png)
|
32 |
|
33 |
+
# **Dialog2Flow single target model** (DSE-base)
|
34 |
|
35 |
This a variation of the **D2F$_{single}$** model introduced in the paper ["Dialog2Flow: Pre-training Soft-Contrastive Action-Driven Sentence Embeddings for Automatic Dialog Flow Extraction"](https://arxiv.org/abs/2410.18481) published in the EMNLP 2024 main conference.
|
36 |
This version uses DSE-base as the backbone model which yields to an increase in performance as compared to the vanilla version using BERT-base as the backbone (results reported in Appendix C).
|