Downloads · 30 days
18
25% of all-time downloads
phuongntc/qwen3-0.6b-legal-rl-lite-mvs-hybrid-n500-k3
qwen3-0.6b-legal-rl-lite-mvs-hybrid-n500-k3 is a machine learning model from phuongntc. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for peft. The card lists the license as apache-2.0.
This repository contains a PEFT/LoRA adapter trained from an SFT adapter using reward-guided pseudo-label selection.
Downloads · 30 days
18
25% of all-time downloads
All-time downloads
73
Public
Repo size
133 MB
Likes
0
Public
Click a slice to open those files.
.pt81.1 MB · 47%
From the Hugging Face model README
This repository contains a PEFT/LoRA adapter trained from an SFT adapter using reward-guided pseudo-label selection.
Qwen/Qwen3-0.6BTask: Vietnamese legal document summarization
Method: RL-lite / reward-guided fine-tuning
Candidate selection: hybrid prompt strategy
Candidate design:
Candidates per document: 3
Filtered training samples: see rl_lite_training_metadata.json
This adapter is intended for research on faithful Vietnamese legal summarization, especially reducing unsupported legal metadata hallucination.
from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig
from peft import PeftModel
import torch
base_model_id = "Qwen/Qwen3-0.6B"
adapter_id = "phuongntc/qwen3-0.6b-legal-rl-lite-mvs-hybrid-n500-k3"
tokenizer = AutoTokenizer.from_pretrained(adapter_id, trust_remote_code=True)
bnb_config = BitsAndBytesConfig(
load_in_4bit=True,
bnb_4bit_quant_type="nf4",
bnb_4bit_compute_dtype=torch.float16,
bnb_4bit_use_double_quant=True,
)
base_model = AutoModelForCausalLM.from_pretrained(
base_model_id,
quantization_config=bnb_config,
device_map="auto",
torch_dtype=torch.float16,
trust_remote_code=True,
)
model = PeftModel.from_pretrained(base_model, adapter_id)
model.eval()