Modelpublic

kanon9999/homework

sourceHugging Faceapache-2.0updated 7mo agoView on Hugging Face

0likes9downloads

Model Card

qwen3-4b-dpo-checkpoint-1500-lora

This repository contains a LoRA adapter checkpoint (step=1500) trained with DPO. Load this adapter on top of the base model.

Base model

—Qwen/Qwen3-4B-Instruct-2507

Usage

python

from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel
import torch

base_model_id = "Qwen/Qwen3-4B-Instruct-2507"
adapter_id = "kanon9999/homework"

tokenizer = AutoTokenizer.from_pretrained(base_model_id)
base_model = AutoModelForCausalLM.from_pretrained(
  base_model_id,
  torch_dtype=torch.float16,
  device_map="auto",
)
model = PeftModel.from_pretrained(base_model, adapter_id)