AksharaLP-MultiIndic-8M / models /kn /training_report.json
sourojitdotcom's picture
Private review: measured five-language release, Hindi sentence mode and custom GGUFs
a449848 verified
Raw History Blame Contribute Delete
567 Bytes
{
"pairs": 3084994,
"counts": {
"aks-Dakshina": 43373,
"aks-Wikidata": 178,
"aks-Samanantar": 36482,
"aks-IndicCorp": 2826694,
"ie": 160660,
"dup_ie": 83571,
"heldout_drop_ie": 1642,
"ie_words": 90000,
"dup_ei": 96771,
"ei": 17607,
"heldout_drop_ei": 100,
"ei_words": 60000
},
"unique_natives": 2904007,
"unique_romans": 2864227,
"overlap_valid_test_natives": 0,
"overlap_valid_test_romans": 0,
"heldout_natives_total": 11062,
"heldout_romans_total": 18017,
"margins": {
"ie": 0.6,
"ei": 0.35,
"maxv_ie": 4,
"maxv_ei": 2
}
}