tokenizer_test3 / recipe /corpus_stats.json
ljcamargo's picture
Upload recipe/corpus_stats.json with huggingface_hub
80724a4 verified
Raw
History Blame Contribute Delete
732 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": 0.2243964589874745,
"old": 0.47716545354444073,
"eng": 0.10020913885448648,
"spa": 0.1002224987104797,
"code": 0.09800644990311855,
"exotic": 0.7015619125319152
},
"bytes": {
"modern": 25933523,
"old": 55146063,
"eng": 11581181,
"spa": 11582725,
"code": 11326616,
"exotic": 81079586
},
"total_bytes": 115570108,
"vocab_target": 256000,
"min_frequency": 2,
"special_token_count": 126,
"language_token_count": 92,
"external_to_exotic_ratio": 0.14285714285714288
}