This repository contains the IMU-1 Base model, a sample-efficient 430M parameter language model introduced in the paper IMU-1: Sample-Efficient Pre-training of Small Language Models.

IMU-1 is trained on 72B tokens and approaches the benchmark performance of models trained on 56× more data.

Model Details

Parameter	Value
Parameters	430M
Hidden dim	1,152
Layers	30
Attention heads	18
KV heads (GQA)	6
Vocab size	49,152
Max context	1,152
Training tokens	72B

Architecture

IMU-1 uses a validated recipe combining recent advances:

QK-norm attention with learnable scale
Per-head gating (sigmoid-based)
Value residual learning
LayerNorm scaling (depth-dependent)
GQA (grouped query attention)
SwiGLU activation
RoPE positional encoding

Training

Optimizer: NorMuon with cautious weight decay, muP parametrization
Schedule: Three-stage WSD (Warmup-Stable-Decay)
Post-processing: Checkpoint EMA (β=0.8)

Usage

python

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained(
    "thepowerfuldeez/imu1_base",
    trust_remote_code=True
)
tokenizer = AutoTokenizer.from_pretrained("thepowerfuldeez/imu1_base")

text = "The quick brown fox"
inputs = tokenizer(text, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=50)
print(tokenizer.decode(outputs[0]))

Note: This model uses custom modeling code. You must pass trust_remote_code=True when loading.

Benchmark Results

Benchmark	Score
HellaSwag (0-shot)	51.1
ARC-Easy	71.4
ARC-Challenge	41.1
PIQA	70.2
Lambada (OpenAI)	51.3
Winograd	74.7
WinoGrande	55.2
BoolQ	59.5
CORE (centered)	30.2

Training Stages

Stage	Iterations	Tokens	Data
1. Stable	100k	29B	DCLM-edu, FineWeb-edu
2. Decay	100k	28B	Higher quality filters
3. Midtrain	65k	14B	Instruction, reasoning, code

Resources

Training Code: sample_efficient_gpt
Stage 1 Data: 1218_imu1_base_stable_corpus
Stage 2 Data: 1226_imu1_base_decay_corpus

Citation

bibtex

@misc{grigorev2026imu1sampleefficientpretrainingsmall,
      title={IMU-1: Sample-Efficient Pre-training of Small Language Models}, 
      author={George Grigorev},
      year={2026},
      eprint={2602.02522},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2602.02522}, 
}

License

Apache 2.0