Datasets:

ArXiv:
License:
orionweller fgaim commited on
Commit
f9ced60
·
verified ·
1 Parent(s): d56af9e

Update README: Fix TiQuAD's language name to Tigrinya (#3)

Browse files

- Update README: Fix TiQuAD's language name to Tigrinya (4688b42d026aba547d972ab0589378f4befa71bd)


Co-authored-by: Fitsum <[email protected]>

Files changed (1) hide show
  1. README.md +2 -2
README.md CHANGED
@@ -40,7 +40,7 @@ This phase dramatically expands language coverage to **1833 languages**, impleme
40
 
41
  - **Temperature Schedule**: τ=0.3 (most uniform sampling)
42
  - **Low-resource Focus**: Includes 1723 new languages with minimal data
43
- - **Rapid Learning**: Demonstrates 68% performance improvement on Tigray and 26% on Faroese
44
  - **Script Diversity**: Covers virtually all writing systems in FineWeb2
45
 
46
  ### Key Innovation: Annealed Language Learning
@@ -68,7 +68,7 @@ Use the script at [this link](https://github.com/JHU-CLSP/mmBERT/blob/main/data/
68
  ## 🎯 Performance Impact
69
 
70
  The decay phase demonstrates remarkable efficiency in low-resource language learning:
71
- - **Tigray (TiQuAD)**: 68% improvement (12.1 F1 points) from including the language
72
  - **Faroese (FoQA)**: 26% improvement (15.4 F1 points)
73
  - **SOTA Performance**: Can even outperforms GPT-4o, Gemini 2.5 Pro
74
  - **Rapid Acquisition**: Significant gains with only 100B tokens of exposure
 
40
 
41
  - **Temperature Schedule**: τ=0.3 (most uniform sampling)
42
  - **Low-resource Focus**: Includes 1723 new languages with minimal data
43
+ - **Rapid Learning**: Demonstrates 68% performance improvement on Tigrinya and 26% on Faroese
44
  - **Script Diversity**: Covers virtually all writing systems in FineWeb2
45
 
46
  ### Key Innovation: Annealed Language Learning
 
68
  ## 🎯 Performance Impact
69
 
70
  The decay phase demonstrates remarkable efficiency in low-resource language learning:
71
+ - **Tigrinya (TiQuAD)**: 68% improvement (12.1 F1 points) from including the language
72
  - **Faroese (FoQA)**: 26% improvement (15.4 F1 points)
73
  - **SOTA Performance**: Can even outperforms GPT-4o, Gemini 2.5 Pro
74
  - **Rapid Acquisition**: Significant gains with only 100B tokens of exposure