ULTRON-Core Standalone Tokenizer (24,576 Vocab)
Pure-Python Byte-Level BPE Tokenizer built from scratch with zero external dependencies.
Specifications
- Vocabulary Size: 24,576 tokens
- Normalization: Unicode NFC Standard
- Byte Fallback: 256-byte leaf tokens (0% Out-of-Vocabulary rate)
- Special Tokens:
<think>,</think>,<dialogue>,<turn>,<user>,<assistant>,<doc>,<para>,<code>,<math>
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support