BloatTokenizer32K / artifacts /tokenizer_config.json
Qvinfor's picture
Upload 4 files
1fdf9f1 verified
Raw History Blame Contribute Delete
2.68 kB
{
"added_vocab_size": 0,
"base_vocab_size": 32000,
"code_ratio": 0.4,
"english_ratio": 0.5,
"eos_token": "<|endoftext|>",
"format_version": 2,
"learned_token_start": 272,
"math_ratio": 0.1,
"merge_scoring": "compression_diversity_v1",
"merges_count": 31728,
"pad_token": "<|pad|>",
"pre_tokenizer": "special_boundaries_whitespace_prefix_v2",
"requested_vocab_size": 32000,
"special_token_count": 16,
"special_token_ids": {
"<|assistant|>": 260,
"<|end_final|>": 267,
"<|end_think|>": 269,
"<|end_tool_call|>": 263,
"<|end_tool_result|>": 265,
"<|end_turn|>": 261,
"<|end_verify|>": 271,
"<|endoftext|>": 256,
"<|final|>": 266,
"<|pad|>": 257,
"<|system|>": 258,
"<|think|>": 268,
"<|tool_call|>": 262,
"<|tool_result|>": 264,
"<|user|>": 259,
"<|verify|>": 270
},
"special_tokens": [
"<|endoftext|>",
"<|pad|>",
"<|system|>",
"<|user|>",
"<|assistant|>",
"<|end_turn|>",
"<|tool_call|>",
"<|end_tool_call|>",
"<|tool_result|>",
"<|end_tool_result|>",
"<|final|>",
"<|end_final|>",
"<|think|>",
"<|end_think|>",
"<|verify|>",
"<|end_verify|>"
],
"stage_budgets": {
"Code": 12691,
"English": 15864,
"Math": 3173
},
"stage_merges": {
"Code": 12691,
"English": 15864,
"Math": 3173
},
"stage_metadata": [
{
"budget": 15864,
"end_token_id_exclusive": 16136,
"learned": 15864,
"rejected_overspecialized_pairs": 18015,
"rejected_redundant_pairs": 0,
"remaining_pairs": 196138,
"remaining_repeated_pairs": 99464,
"stage": "English",
"start_token_id": 272,
"stop_reason": "budget_filled",
"unfilled": 0
},
{
"budget": 3173,
"end_token_id_exclusive": 19309,
"learned": 3173,
"rejected_overspecialized_pairs": 5609,
"rejected_redundant_pairs": 0,
"remaining_pairs": 57739,
"remaining_repeated_pairs": 29346,
"stage": "Math",
"start_token_id": 16136,
"stop_reason": "budget_filled",
"unfilled": 0
},
{
"budget": 12691,
"end_token_id_exclusive": 32000,
"learned": 12691,
"rejected_overspecialized_pairs": 29284,
"rejected_redundant_pairs": 0,
"remaining_pairs": 430310,
"remaining_repeated_pairs": 183753,
"stage": "Code",
"start_token_id": 19309,
"stop_reason": "budget_filled",
"unfilled": 0
}
],
"stage_unfilled_slots": {
"Code": 0,
"English": 0,
"Math": 0
},
"strict_curriculum": true,
"tokenizer_name": "bloat-tokenizer",
"vocab_size": 32000
}