---
library_name: peft
tags:
- lora
- qwen
- emotion-support
base_model: Qwen/Qwen3.5-9B
pipeline_tag: text-generation
---

# Emotion Support

LoRA adapters fine-tuned on **Qwen/Qwen3.5-9B** for emotion support.

## Subfolders

| Folder | Description |
|--------|-------------|
| `sft-lora` | SFT LoRA |
| `grpo-lora` | GRPO LoRA |
| `grpo-lora-shot2` | GRPO LoRA (shot2) |
| `grpo-lora-shot3` | GRPO LoRA (shot3) |

## Load

```python
from peft import PeftModel
from transformers import AutoModelForCausalLM, AutoTokenizer

base_id = "Qwen/Qwen3.5-9B"
model = AutoModelForCausalLM.from_pretrained(base_id, torch_dtype="auto", device_map="auto")
model = PeftModel.from_pretrained(model, "ForestHYS/Emotion-Support", subfolder="grpo-lora-shot3")
tokenizer = AutoTokenizer.from_pretrained("ForestHYS/Emotion-Support", subfolder="grpo-lora-shot3")
```