pico-test / tokenizer_training_manifest.json
Banaxi-Tech's picture
Publish BananaMind 2 Pico 384-token tokenizer
4c8a7d2 verified
Raw
History Blame Contribute Delete
767 Bytes
{
"format": "bananamind2_pico_tokenizer_v1",
"stage": "tokenizer_complete",
"vocab_size": 384,
"model_max_length": 8192,
"target_gib": 1.0,
"sources": [
{
"name": "fineweb_hq",
"dataset_id": "epfml/FineWeb-HQ",
"config": null,
"text_field": "text",
"byte_share": 0.75,
"revision": "e58199cdd52438d94405df1a4d8630cc5f13bf84",
"target_bytes": 805306368,
"consumed_bytes": 805307468
},
{
"name": "cosmopedia_v2",
"dataset_id": "HuggingFaceTB/smollm-corpus",
"config": "cosmopedia-v2",
"text_field": "text",
"byte_share": 0.25,
"revision": "3ba9d605774198c5868892d7a8deda78031a781f",
"target_bytes": 268435456,
"consumed_bytes": 268438475
}
]
}