Instructions to use Salesforce/RetailSim-GRPO-4B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use Salesforce/RetailSim-GRPO-4B with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("google/gemma-3-4b-it") model = PeftModel.from_pretrained(base_model, "Salesforce/RetailSim-GRPO-4B") - Notebooks
- Google Colab
- Kaggle
File size: 425 Bytes
17f387d 4932dc3 17f387d 4932dc3 17f387d 4932dc3 17f387d 4932dc3 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 | ---
base_model: google/gemma-3-4b-it
library_name: peft
tags:
- gemma3
- lora
- peft
- vllm
---
# UserGRPO Model for Retail Simulations
UserGRPO checkpoint based on Gemma-3 for agent simulations in [SalesSim](https://arxiv.org/html/2605.08334v1). This model has been trained to adhere to persona specifications in dialogue-based retail simulations, as specified by preferences and dealbreakers.
## License
CC-BY-NC 4.0
|