Benchmark Languages

benchmarkLanguages component contains data of Resources (MALA), Writing Systems, Resources (Joshi).

Data Structure:

• Each Benchmark contains Resources (MALA), Writing Systems and Resources (Joshi).

• Under Resources (MALA), Languages are divided into High, Medium-High, Medium, Medium-Low, Low, Unseen.

• Under Resources (Joshi), Languages are divided into Category 0, Category 1, Category 2, Category 3, Category 4 and Category 5.

• Under Writing Systems, Languages are divided according to their writing systems

Example Structure:

export const benchmarkLanguages = {
  "SIB-200": {
    ResourcesMALA: {
      High: ["Avg", "tgk_Cyrl", "guj_Gujr", ...],
      "Medium-High": ["Avg", "amh_Ethi", "ast_Latn", ...],
      Medium: ["Avg", "fra_Latn", "glg_Latn", ...],
      "Medium-Low": ["Avg", "ace_Latn", "aka_Latn", ...],
      Low: ["Avg", "acm_Arab", "ajp_Arab"],
      Unseen: ["Avg", "ace_Arab", "acq_Arab"]
    },

    writingSystems: {
      Arabic: ["ace_Arab", ...],
      ...
    },

    ResourcesJoshi: {

       "Category 0": ["umb_Latn", ...],
       "Category 1": ["bel_Cyrl", ...],
       "Category 2": ["zul_Latn", ...],
       "Category 3": ["zsm_Latn", ...],
       "Category 4": ["vie_Latn", ...],
       "Category 5": ["spa_Latn", ...]

    }

  },

  ...

}