Persim Gemma 12b
An open-weight Big-Five personality → household layout generator for the PerSim pipeline (arXiv:2607.00022, IROS 2026). Given the pipeline's generatelayouts.py prompt (scene + persona + item vocabulary), it produces...
Fonte do modelo
- Tipo
- Other
- Tamanho do arquivo
- 3,84–4,64 GB
- Modelo base
- google/gemma-4-12B-it
- Versão
- vRevision ccb197e
- Atualizado
- 14 de set. de 2026
- Downloads
- 925
- Fonte
- Hugging Face
Descrição da fonte
persim-gemma-12b
An open-weight Big-Five personality → household layout generator for the PerSim pipeline (arXiv:2607.00022, IROS 2026). Given the pipeline's generate_layouts.py prompt (scene + persona + item vocabulary), it produces personality-conditioned placements as strict JSON:
{
"room_movable_items": {"kitchen_0": ["mug_0", "plate_0", ...], ...},
"initial_object_positions": {"mug_0": {"relation": "on", "anchor_object": "countertop_tpuwys_0"}, ...}
}
The pipeline works with any OpenAI-compatible endpoint; this model is a ready-made open option for the layout stage (the paper's fine-tuned Gemini 2.5 Flash cannot be redistributed). Outputs inherit only the pipeline's automatic checks — no plausibility or paper-equivalence claims.
Training
QLoRA (r=16) on google/gemma-4-12B-it, 2 epochs on one RTX 4090. Data: persim-sft human placement anchors (rigidity dropped) + layout demonstrations self-generated by the base model on the real pipeline prompts and filtered through check_layouts.py. No proprietary-model outputs in the supervision.
Usage
vllm serve xianyao/persim-gemma-12b \
--max-model-len 16384 \
--override-generation-config '{"repetition_penalty": 1.05}'
python generate_persona.py --model google/gemma-4-12B-it # base model
python generate_layouts.py --model xianyao/persim-gemma-12b # this model
python generate_trajectories.py --model google/gemma-4-12B-it # base model
Also answers the standalone anchor query (Personality: O=..., C=...\nObject: mug → {"rooms": [...], "cooccurrence": [...]}).
Limitations
- The largest shipped scene (
house_single_floor, 120–240 items) is unreliable: generation can run away into numbered duplicates until the token limit.repetition_penalty=1.05plus the pipeline's retries partially compensate. Smaller scenes are stable. - Trained on the pipeline's five scenes and item vocabulary; new scenes are untested.
- Layout plausibility is not validated — format and containment checks only.
Citation
@article{li2026personalize,
title = {When to Personalize Household Object Search: A Rigidity-Gated Hybrid Policy},
author = {Li, Xianyao and Wang, Yuhai and Xiao, Hu and Smith, Kaleb and Ye, Gilbert Yang and Du, Eric Jing},
journal = {arXiv preprint arXiv:2607.00022},
year = {2026}
}
Derived from Gemma. Gemma is provided by Google; use is subject to the base model's license and terms.
Fontes
1 fonte
Fontes
1 fonteVerificado 9 de ago.
Artefatos de modelo
5 artefatos
Artefatos de modelo
5 artefatosmodel-00001-of-00005.safetensors
safetensors · 4,62 GB · SHA-256 e153bc61efcd…a731 · Hugging Face
Baixarmodel-00002-of-00005.safetensors
safetensors · 4,63 GB · SHA-256 15db0f190a89…ce59 · Hugging Face
Baixarmodel-00003-of-00005.safetensors
safetensors · 4,55 GB · SHA-256 0c67c30db15b…cb0f · Hugging Face
Baixarmodel-00004-of-00005.safetensors
safetensors · 4,64 GB · SHA-256 2da3ad85d04b…25a0 · Hugging Face
Baixarmodel-00005-of-00005.safetensors
safetensors · 3,84 GB · SHA-256 0c8be1660790…d6cf · Hugging Face
BaixarTrechos de fonte
3 trechos
Trechos de fonte
3 trechos--- license: apache-2.0 base_model: google/gemma-4-12B-it datasets: - xianyao/persim-sft language: - en library_name: transformers pipeline_tag: text-generation tags: - robotics - household - personality - layout-generation - gemma - lora --- # persim-gemma-12b An open-weight **Big-Five personality → household layout** generator for the [PerSim](https://github.com/XianyaoLi/PerSim) pipeline ([arXiv:2607.00022](https://arxiv.org/abs/2607.00022), IROS 2026). Given the pipeline's `generate_layouts.py` prompt (scene + persona + item vocabulary), it produces personality-conditioned placements as strict JSON: ```json { "room_movable_items": {"kitchen_0": ["mug_0", "plate_0", ...], ...}, "initial_object_positions": {"mug_0": {"relation": "on", "anchor_object": "countertop_tpuwys_0"}, ...} } ``` The pipeline works with any OpenAI-compatible endpoint; this model is a ready-made open option for the layout stage (the paper's fine-tuned Gemini 2.5 Flash cannot be redistributed). Outputs inherit only the pipeline's automatic checks — no plausibility or paper-equivalence claims. ## Training QLoRA (r=16) on `google/gemma-4-12B-it`, 2 epochs on one RTX 4090. Data: [persim-sft](https://huggingface.co/datasets/xianyao/persim-sft) human placement anchors (rigidity dropped) + layout demonstrations self-generated by the base model on the real pipeline prompts and filtered through `check_layouts.py`. No proprietary-model outputs in the supervision. ## Usage ```bash vllm serve xianyao/persim-gemma-12b \ --max-model-len 16384 \ --override-generation-config '{"repetition_penalty": 1.05}' ``` ```bash python generate_persona.py --model google/gemma-4-12B-it # base model python generate_layouts.py --model xianyao/persim-gemma-12b # this model python generate_trajectories.py --model google/gemma-4-12B-it # base model ``` Also answers the standalone anchor query (`Personality: O=..., C=...\nObject: mug` → `{"rooms": [...], "cooccurrence": [...]}`). ## Limitations - The largest shipped scene (`house_single_floor`, 120–240 items) is unreliable: generation can run away into numbered duplicates until the token limit. `repetition_penalty=1.05` plus the pipeline's retries partially compensate. Smaller scenes are stable. - Trained on the pipeline's five scenes and item vocabulary; new scenes are untested. - Layout plausibility is not validated — format and containment checks...
Source context: 925 downloads · 0 likes · Pipeline text-generation · Library transformers · Repo xianyao/persim-gemma-12b
Source context: 1287 downloads · 0 likes · Pipeline text-generation · Library transformers · Repo xianyao/persim-gemma-12b
