Packages

Fast Hugging Face tokenizer.json, OpenAI tiktoken, and SentencePiece model bindings for Elixir via the IREE tokenizer runtime

Current section

Files

Jump to
iree_tokenizers test fixtures minimal_unigram.json
Raw

test/fixtures/minimal_unigram.json

{
"version": "1.0",
"model": {
"type": "Unigram",
"vocab": [
["<unk>", -10.0],
["▁hello", -1.0],
["▁world", -1.0]
],
"unk_id": 0
},
"pre_tokenizer": {
"type": "Metaspace",
"replacement": "▁",
"prepend_scheme": "always",
"split": true
},
"decoder": {
"type": "Metaspace",
"replacement": "▁",
"prepend_scheme": "always",
"split": true
}
}