max-0.1b v1

A small Korean-first base language model (109.5M parameters) for industrial small functions such as document classification, request routing, field extraction and form filling. v1 continues v0 from its pre-decay checkpoint for 7.6B more tokens (10.0B in total). This is a base model (not instruction-tuned): a starting point for task fine-tuning and research, not for direct end-user use.

ํ•œ๊ตญ์–ด ์ค‘์‹ฌ์˜ ์†Œํ˜• ๊ธฐ์ดˆ ์–ธ์–ด๋ชจ๋ธ(1.095์–ต ๋งค๊ฐœ๋ณ€์ˆ˜)์ž…๋‹ˆ๋‹ค. ๋ฌธ์„œ ๋ถ„๋ฅ˜, ์š”์ฒญ ๋ถ„๊ธฐ, ํ•ญ๋ชฉ ์ถ”์ถœ, ์–‘์‹ ์ž‘์„ฑ ๊ฐ™์€ ์‚ฐ์—…์šฉ ์ž‘์€ ๊ธฐ๋Šฅ์˜ ์ถœ๋ฐœ์ ์œผ๋กœ ์“ฐ๋Š” ๊ธฐ์ดˆ ๋ชจ๋ธ์ž…๋‹ˆ๋‹ค. ์ง€์‹œ ๋”ฐ๋ฅด๊ธฐ ๋ฏธ์„ธ์กฐ์ •์€ ํ•˜์ง€ ์•Š์•˜์Šต๋‹ˆ๋‹ค. v0์˜ ๊ฐ์‡  ์ „ ์ €์žฅ๋ถ„์—์„œ ์ด์–ด์„œ ํ•™์Šตํ–ˆ๊ณ , ๋ˆ„์  ํ•™์Šต๋Ÿ‰์€ 100์–ต ํ† ํฐ์ž…๋‹ˆ๋‹ค.

Files

File Use
model.safetensors, config.json, tokenizer.json Transformers (Qwen3ForCausalLM-compatible layout)
gguf/max01b-v1-f16.gguf (221 MB) llama.cpp, full precision (greedy output identical to Transformers on 5/5 test prompts)
gguf/max01b-v1-q4_k_m.gguf (73 MB) llama.cpp, 4-bit, runs on ordinary CPUs (perplexity about 8% higher than f16)

Context length: 1,024 tokens. Vocabulary: 32,000 (Korean-optimized, digits split; same tokenizer as v0). Input text should be Unicode NFC-normalized (NFD Korean inflates the token count about 10x).

llama-completion -m max01b-v1-q4_k_m.gguf -p "์˜ค๋Š˜ ํšŒ์˜์—์„œ ๋…ผ์˜๋œ ์•ˆ๊ฑด์€" -n 64 -no-cnv
from transformers import AutoModelForCausalLM, PreTrainedTokenizerFast
tok = PreTrainedTokenizerFast.from_pretrained("quantumaikr/max-0.1b-v1")
model = AutoModelForCausalLM.from_pretrained("quantumaikr/max-0.1b-v1")

Training

  • Architecture as v0: 12 layers, width 768, SwiGLU 2048, RoPE, QK-norm, tied embeddings (Qwen3-compatible names).
  • Continued from v0 stable_pre_decay (2.40B tokens) for 7.60B tokens on 2 GPUs (DDP); last 20% linear learning-rate decay.
  • Optimizer: Muon (hidden matrices) + AdamW (embeddings, scalars). Batch 524,288 tokens, context 1,024.
  • Data (all ODC-By 1.0): FineWeb-2 Korean, FinePDFs Korean, FineWeb-Edu (incl. score โ‰ฅ 4 subset), FineMath-4plus. Mix: Korean web 40% / Korean documents 20% / English educational 33% / math 7% (decay phase: 33 / 25 / 32 / 10). Cleaning: NFC, script-ratio and length filters, exact deduplication, regex replacement of e-mail, phone, account and resident-registration-number patterns. No code data, no B/C-grade corpora, no model-generated text.

Evaluation

Validation loss (held-out, same files as v0; lower is better):

set v0 (2.8B tokens) v1 (10.0B)
Korean web 3.199 3.055 (0.903 bpb)
English 2.961 2.851 (1.020 bpb)
Korean documents (PDF) 2.341 2.208 (0.791 bpb)
Math 2.185 2.020 (0.936 bpb)

Fine-tuned on public, human-written Korean tasks (train split, 2 epochs; LR chosen on dev; test mean ยฑ std over 3 seeds):

task v0 v1 KLUE-RoBERTa-base (ref.)
KLUE-YNAT (macro F1) 84.4 ยฑ 0.3 85.1 ยฑ 0.0 86.4
KLUE-NLI (acc) 70.4 ยฑ 0.6 73.8 ยฑ 1.0 85.1
KLUE-NER (entity F1) 81.1 ยฑ 0.3 82.5 ยฑ 0.4 87.4
NSMC (acc) 86.4 87.1 ยฑ 0.1 88.4
MASSIVE ko intent (acc) 86.4 ยฑ 0.2 86.1 ยฑ 0.4 86.5
MASSIVE ko slot (F1) 74.6 ยฑ 0.1 74.9 ยฑ 0.3 74.6
3i4K (acc) 89.6 ยฑ 0.0 89.5 ยฑ 0.3 90.0

The decoder is generative (answers are generated text scored by the same scorer for every model); the encoder reference uses classification/tagging heads on the same splits.

Limitations

  • Small base model: limited world knowledge and reasoning; it can produce fluent but wrong text. Fine-tune before use.
  • Personal-information patterns were replaced with regular expressions; names and other free-form personal data may remain in the training text. Do not use the model to extract personal data.
  • Korean and English only; context 1,024 tokens.

License

Apache-2.0 for the model weights. Training data: FineWeb-2, FinePDFs, FineWeb-Edu and FineMath (Hugging Face), ODC-By 1.0; please keep the attribution when redistributing derived datasets.

Downloads last month
125
Safetensors
Model size
0.1B params
Tensor type
BF16
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support