Hugging Face's logo

Flaglab
/
SciBETO-base

Fill-Mask
Transformers
PyTorch
Safetensors
Spanish
roberta
spanish
scientific
Model card Files
xet
Community
2
juanks4321 commited on
Commit
3f05e05
·
verified ·
1 Parent(s): 0d1d2a6

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +11 -10
README.md CHANGED
@@ -67,16 +67,17 @@ The final dataset covers multiple disciplines (economics, medicine, engineering,
67
 
68
  Sci-BETO was fine-tuned and benchmarked across multiple downstream tasks, both general-domain and scientific:
69
 
70
- | **Dataset** | **Metric** | **Sci-BETO Large** | **Sci-BETO Base** | **BETO** | **BERTIN** |
71
- |--------------|-------------|-------------------:|------------------:|----------:|------------:||
72
- | **WikiCAT** | F1 (macro) | **0.7738** | 0.7583 | 0.7624 | 0.7598 |
73
- | **PAWS-X (es)** | F1 (macro) | **0.9148** | 0.8794 | 0.8985 | 0.8961 |
74
- | **PharmaCoNER** | F1 (micro) | **0.8959** | 0.8733 | 0.8845 | 0.8802 |
75
- | **CANTEMIST** | F1 (micro) | 0.8809 | 0.8784 | 0.8954 | **0.8956** |
76
- | **NLI (ESNLI-R)** | F1 (micro) | — | — | — | — |
77
- | **BanRep (JEL)** | Exact Match | **0.6116** | 0.6043 | 0.5933 | 0.5807 |
78
- | **Rosario** | F1 (macro) | **0.9203** | 0.9194 | 0.9079 | 0.9121 |
79
- | **Econ-IE** | F1 (micro) | **0.5256** | 0.5158 | 0.5199 | 0.4992 |
 
80
 
81
  On average, **Sci-BETO** achieves comparable or superior results to general-domain Spanish models in specialized contexts (scientific, biomedical, economic), while maintaining strong performance in general text understanding.
82
 
 
67
 
68
  Sci-BETO was fine-tuned and benchmarked across multiple downstream tasks, both general-domain and scientific:
69
 
70
+ | **Dataset** | **Metric** | **Sci-BETO Large** | **Sci-BETO Base** | **BETO** | **BERTIN** |
71
+ |---------------------|----------------|-------------------:|------------------:|----------:|------------:|
72
+ | **WikiCAT** | F1 (macro) | **0.7738** | 0.7583 | 0.7624 | 0.7598 |
73
+ | **PAWS-X (es)** | F1 (macro) | **0.9148** | 0.8794 | 0.8985 | 0.8961 |
74
+ | **PharmaCoNER** | F1 (micro) | **0.8959** | 0.8733 | 0.8845 | 0.8802 |
75
+ | **CANTEMIST** | F1 (micro) | 0.8809 | 0.8784 | 0.8954 | **0.8956** |
76
+ | **NLI (ESNLI-R)** | F1 (micro) | — | — | — | — |
77
+ | **BanRep (JEL)** | Exact Match | **0.6116** | 0.6043 | 0.5933 | 0.5807 |
78
+ | **Rosario** | F1 (macro) | **0.9203** | 0.9194 | 0.9079 | 0.9121 |
79
+ | **Econ-IE** | F1 (micro) | **0.5256** | 0.5158 | 0.5199 | 0.4992 |
80
+
81
 
82
  On average, **Sci-BETO** achieves comparable or superior results to general-domain Spanish models in specialized contexts (scientific, biomedical, economic), while maintaining strong performance in general text understanding.
83