tokenizer_test4 / recipe /corpus_stats.json
ljcamargo's picture
Upload recipe/corpus_stats.json with huggingface_hub
387cbb7 verified
Raw
History Blame Contribute Delete
738 Bytes
{
"recipe_version": "hf-bytelevel-bpe-v2-iterator-eval",
"test_run": true,
"weights_target": {
"exotic": 0.7,
"eng": 0.1,
"spa": 0.1,
"code": 0.1
},
"weights_actual": {
"modern": 0.30920752033535065,
"old": 0.39112290962568264,
"eng": 0.10004712984655464,
"spa": 0.10002112797536222,
"code": 0.09960131221704985,
"exotic": 0.7003304299610332
},
"bytes": {
"modern": 111568315,
"old": 141125041,
"eng": 36099024,
"spa": 36089642,
"code": 35938164,
"exotic": 252693356
},
"total_bytes": 360820186,
"vocab_target": 256000,
"min_frequency": 2,
"special_token_count": 274,
"language_token_count": 240,
"external_to_exotic_ratio": 0.14285714285714288
}