LoRA fine-tune of unsloth/gemma-3-270m-it specialized for the ALCF Aurora supercomputer (Intel Xeon Sapphire Rapids + Intel GPU Max 1550 / Ponte Vecchio, oneAPI / SYCL, PBS Pro).

Off-the-shelf code-LLMs hallucinate Aurora specifics — they suggest nvcc instead of icpx -fsycl, srun / aprun instead of mpiexec, NERSC's /global/cfs instead of /lus/flare, and CUDA device strings instead of xpu. This adapter teaches the base model the actual Aurora toolchain, file system layout, scheduler conventions, and recommended PyTorch/TensorFlow/SYCL idioms.

Model summary


Base model	`unsloth/gemma-3-270m-it`
Format	GGUF, f16 — single file, llama.cpp / Ollama / LM Studio compatible
Fine-tuning	LoRA (PEFT) — r=32, α=64, dropout 0.0, 2 epochs
Optimizer	AdamW fused, lr 2e-4 cosine, warmup 3%, batch 1 × grad-accum 8
Precision / seq-len	bf16, 1,536 tokens
Training data	`aurora-docs-distill-v2-datascience` — 1,117 ChatML rows
Train loss (final)	1.2462
Hardware	1 Aurora PVC tile (1/12 of a node, 64 GB HBM), IPEX + PyTorch 2.10 XPU backend
Eval (53-Q Aurora, 0–5)	pending

Quick start

On Aurora (PVC GPU, SYCL llama.cpp build) — interactive PBS session:

bash

# 1. Grab a debug node
qsub -I -A <project> -q debug -l select=1,walltime=01:00:00,filesystems=home:flare

# 2. Load the toolchain
module load frameworks
source /lus/flare/projects/<project>/scripts/env.sh   # or your own oneAPI setup
export ONEAPI_DEVICE_SELECTOR=level_zero:gpu

# 3. Download to flare (NOT $HOME — quota is small)
hf download shazzadulimun/gemma3-270m-aurora-ml-v3-gguf --local-dir /lus/flare/projects/<project>/models/aurora-chat-v3

# 4. Run on a single PVC tile
/path/to/llama.cpp/build_sycl/bin/llama-cli \
    -m /lus/flare/projects/<project>/models/aurora-chat-v3/*.gguf \
    -ngl 999 -sm none --temp 0.0 -cnv \
    -p "How do I launch one MPI rank per GPU tile on Aurora?"

Anywhere else (laptop, workstation, any GPU):

bash

hf download shazzadulimun/gemma3-270m-aurora-ml-v3-gguf --local-dir ./model
./llama-cli -m ./model/*.gguf -ngl 999 --temp 0.0 -cnv

Or Ollama / LM Studio: ollama run hf.co/shazzadulimun/gemma3-270m-aurora-ml-v3-gguf

Training data

Distilled from openai/gpt-oss-120b on ALCF Sophia (vLLM) over 130 cleaned chunks of docs.alcf.anl.gov/aurora. 1,117 training rows + 139 validation rows in ChatML format with embedded chain-of-thought (**Reasoning:** / **Answer:**).

Topic specialist — Data Science / AI. Subset filtered to PyTorch/XPU, TensorFlow, vLLM, DeepSpeed, Megatron, JAX, Jupyter, and ML framework setup on Aurora. Trains a model to know torch.xpu, module load frameworks, IPEX patterns, etc.

Full corpus + reproduction scripts: SIslamMun/Generator @ aurora-datasets-2026-04-30.

Limitations

Synthetic-data biases. Teacher (gpt-oss-120b) can confabulate plausible-looking but incorrect commands. Treat outputs as a verifiable first draft, not authoritative.
Doc snapshot is fixed at 2026-04-29. Module versions, queue names, and APIs change — anything published after that date isn't reflected here.
Aurora-only. Specifics (/lus/flare, xpu, PBS queues) won't transfer to Frontier, Polaris, or other systems.
Use temperature ≤ 0.1 for technical answers; higher temps invite invented flag names and paths.

Citation

bibtex

@misc{aurora-llms-2026,
  title  = { Gemma-3-270M-Aurora-ML v3 },
  author = { Islam Mun, Shazzadul },
  year   = { 2026 },
  url    = { https://huggingface.co/shazzadulimun/gemma3-270m-aurora-ml-v3-gguf },
  note   = { LoRA fine-tune of gemma-3-270m-it; data distilled from gpt-oss-120b on docs.alcf.anl.gov/aurora }
}

License

Apache-2.0 for the adapter weights and synthetic training data. Source corpus is public ALCF user documentation. Base model retains its own license — see unsloth/gemma-3-270m-it.