from peft import PeftModel, AutoModelForCausalLM, AutoTokenizer
import torch

base = AutoModelForCausalLM.from_pretrained(
    "mistralai/Mistral-7B-v0.3",
    torch_dtype=torch.float16,
    device_map="auto"
)
model = PeftModel.from_pretrained(base, "Sharpener9290/hcup-llm-humanizer-v1")
tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-v0.3")

Training Data

400 instruction pairs from HCUP corpus (nis, neds, trinetx, hcup_general)
50 preference pairs for DPO humanization style transfer

LoRA Config

rank=16, alpha=32
target_modules: q_proj, v_proj
dropout=0.05