rafmacalaba commited on
Commit
d2803e8
·
verified ·
1 Parent(s): 37b2807

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +23 -0
README.md CHANGED
@@ -40,3 +40,26 @@ Fine-tune of `urchade/gliner_large-v2.1` for data-use mention extraction
40
  **Best F0.5**: 0.8506 (thr=0.6)
41
  **Best F1**: 0.8549 (thr=0.5)
42
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
40
  **Best F0.5**: 0.8506 (thr=0.6)
41
  **Best F1**: 0.8549 (thr=0.5)
42
 
43
+ <!-- NER_COMPARISON_START -->
44
+ ## NER holdout comparison
45
+ device: NVIDIA H100 NVL
46
+
47
+ ### rafmacalaba/data-use-mentions-extended (n=9249)
48
+
49
+ | model | backend | best F0.5 | thr | best F1 | thr | wall-clock (s) | texts/s |
50
+ | --- | --- | --- | --- | --- | --- | --- | --- |
51
+ | `rafmacalaba/gliner_datause_extended` | gliner | 0.8506 | 0.6 | 0.8549 | 0.5 | 202.7 | 45.6 |
52
+ | `ai4data/gliner2_datause` | gliner2 | 0.8634 | 0.7 | 0.8624 | 0.6 | 180.8 | 51.1 |
53
+
54
+ F0.5 by threshold (sweet spots side-by-side):
55
+
56
+ | thr | `rafmacalaba/gliner_datause_extended` | `ai4data/gliner2_datause` |
57
+ | --- | --- | --- |
58
+ | 0.1 | 0.6476 | 0.7321 |
59
+ | 0.2 | 0.7121 | 0.7712 |
60
+ | 0.3 | 0.7505 | 0.7979 |
61
+ | 0.4 | 0.7858 | 0.8201 |
62
+ | 0.5 | 0.8224 | 0.8363 |
63
+ | 0.6 | 0.8506 | 0.8523 |
64
+ | 0.7 | 0.8422 | 0.8634 |
65
+ <!-- NER_COMPARISON_END -->