Hugging Face's logo

rafmacalaba
/
gliner_datause

Token Classification
GLiNER
PyTorch
ner
data-use
Model card Files
xet
 Community
rafmacalaba commited on
Commit
c4bf8a6
·
verified ·
1 Parent(s): ed62505

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +23 -11
README.md CHANGED
@@ -20,6 +20,7 @@ Fine-tune of `urchade/gliner_large-v2.1` for data-use mention extraction
20
  ## Training
21
  - base model: `urchade/gliner_large-v2.1`
22
  - dataset: `rafmacalaba/data-use-mentions` (gliner config)
 
23
  - epochs: 5
24
  - learning rate: 5e-06
25
  - batch size: 16
@@ -29,14 +30,25 @@ Fine-tune of `urchade/gliner_large-v2.1` for data-use mention extraction
29
 
30
  | thr | tp | fp | fn | precision | recall | f0.5 | f1 |
31
  | --- | --- | --- | --- | --- | --- | --- | --- |
32
- | 0.10 | 7147 | 5212 | 199 | 0.5783 | 0.9729 | 0.6293 | 0.7254 |
33
- | 0.20 | 7082 | 3928 | 264 | 0.6432 | 0.9641 | 0.6891 | 0.7716 |
34
- | 0.30 | 7017 | 3234 | 329 | 0.6845 | 0.9552 | 0.7256 | 0.7975 |
35
- | 0.40 | 6901 | 2549 | 445 | 0.7303 | 0.9394 | 0.7643 | 0.8217 |
36
- | 0.50 | 6704 | 1874 | 642 | 0.7815 | 0.9126 | 0.8046 | 0.8420 |
37
- | 0.60 | 6212 | 1196 | 1134 | 0.8386 | 0.8456 | 0.8400 | 0.8421 |
38
- | 0.70 | 4923 | 645 | 2423 | 0.8842 | 0.6702 | 0.8311 | 0.7624 |
39
-
40
- **Best F0.5**: 0.8400 (thr=0.6)
41
- **Best F1**: 0.8421 (thr=0.6)
42
-
 
 
 
 
 
 
 
 
 
 
 
 
20
  ## Training
21
  - base model: `urchade/gliner_large-v2.1`
22
  - dataset: `rafmacalaba/data-use-mentions` (gliner config)
23
+ - corpus: `all`
24
  - epochs: 5
25
  - learning rate: 5e-06
26
  - batch size: 16
 
30
 
31
  | thr | tp | fp | fn | precision | recall | f0.5 | f1 |
32
  | --- | --- | --- | --- | --- | --- | --- | --- |
33
+ | 0.10 | 16046 | 9301 | 123 | 0.6331 | 0.9924 | 0.6825 | 0.7730 |
34
+ | 0.20 | 16016 | 6724 | 153 | 0.7043 | 0.9905 | 0.7475 | 0.8233 |
35
+ | 0.30 | 15971 | 5288 | 198 | 0.7513 | 0.9878 | 0.7890 | 0.8534 |
36
+ | 0.40 | 15895 | 4053 | 274 | 0.7968 | 0.9831 | 0.8282 | 0.8802 |
37
+ | 0.50 | 15694 | 2871 | 475 | 0.8454 | 0.9706 | 0.8678 | 0.9037 |
38
+ | 0.60 | 14185 | 1753 | 1984 | 0.8900 | 0.8773 | 0.8874 | 0.8836 |
39
+ | 0.70 | 11161 | 827 | 5008 | 0.9310 | 0.6903 | 0.8703 | 0.7928 |
40
+
41
+ **Best F0.5**: 0.8874 (thr=0.6)
42
+ **Best F1**: 0.9037 (thr=0.5)
43
+
44
+ ## Evaluation breakdown (holdout)
45
+
46
+ | group | examples | spans | thr | precision | recall | f0.5 | f1 |
47
+ | --- | ---: | ---: | ---: | ---: | ---: | ---: | ---: |
48
+ | overall | 12531 | 16417 | 0.60 | 0.8900 | 0.8773 | 0.8874 | 0.8836 |
49
+ | prwp | 9079 | 12380 | 0.60 | 0.8886 | 0.8935 | 0.8896 | 0.8911 |
50
+ | fcv | 3452 | 4037 | 0.50 | 0.8670 | 1.0000 | 0.8907 | 0.9287 |
51
+ | general_prwp | 9079 | 12380 | 0.60 | 0.8886 | 0.8935 | 0.8896 | 0.8911 |
52
+ | fcv_pads_east_asia | 784 | 863 | 0.50 | 0.8877 | 1.0000 | 0.9081 | 0.9405 |
53
+ | jdc_operational | 163 | 178 | 0.60 | 0.9193 | 0.8555 | 0.9058 | 0.8862 |
54
+ | refugee_pads | 803 | 875 | 0.50 | 0.8475 | 1.0000 | 0.8742 | 0.9175 |