You can deploy this model and start earning money today!

jhsu12/
solidity-vuln-cls-integer-overflow-underflow-v1

Name: solidity-vuln-cls-integer-overflow-underflow-v1
Availability: InStock

peft

safetensors

lora

sequence-classification

solidity

smart-contract

vulnerability-detection

integer-overflow-underflow

text-classification

Qwen/Qwen2.5-Coder-3B-Instruct

adapter

apache-2.0

Views

No views yet

Solidity Vulnerability Classifier — Integer Overflow/Underflow

Binary classifier that detects Integer Overflow/Underflow vulnerabilities in Solidity smart contracts.

Model Details

Base model: Qwen/Qwen2.5-Coder-3B-Instruct
Method: QLoRA (4-bit NF4) + classification head
Task: Sequence Classification (2 labels: safe / vulnerable)
LoRA rank: 16, targeting q_proj, k_proj, v_proj, o_proj
Classification head: modules_to_save=["score"]

Available Checkpoints

Load a specific checkpoint with revision=:

model = PeftModel.from_pretrained(base, "jhsu12/solidity-vuln-cls-integer-overflow-underflow-v1", revision="checkpoint-200")

Tag	Step
`checkpoint-316`	316
`checkpoint-632`	632
`checkpoint-948`	948
`checkpoint-1264`	1264
`checkpoint-1580`	1580 ← `main`

Usage

python

from transformers import AutoTokenizer, AutoModelForSequenceClassification, BitsAndBytesConfig
from peft import PeftModel
import torch

base_model = "Qwen/Qwen2.5-Coder-3B-Instruct"
bnb_config = BitsAndBytesConfig(load_in_4bit=True, bnb_4bit_quant_type="nf4",
    bnb_4bit_compute_dtype=torch.bfloat16, bnb_4bit_use_double_quant=True)

model = AutoModelForSequenceClassification.from_pretrained(
    base_model, num_labels=2, quantization_config=bnb_config,
    device_map="auto", trust_remote_code=True, ignore_mismatched_sizes=True)
model = PeftModel.from_pretrained(model, "jhsu12/solidity-vuln-cls-integer-overflow-underflow-v1")
model.eval()

tokenizer = AutoTokenizer.from_pretrained("jhsu12/solidity-vuln-cls-integer-overflow-underflow-v1", trust_remote_code=True)

code = "pragma solidity ^0.8.0; contract Example { ... }"
inputs = tokenizer(code, return_tensors="pt", truncation=True, max_length=1536).to(model.device)
with torch.no_grad():
    logits = model(**inputs).logits
probs = torch.softmax(logits, dim=-1)
print(f"Safe: {probs[0][0]:.2%}, Vulnerable: {probs[0][1]:.2%}")

Or use the inference script:

python inference_classifier.py --checkpoint jhsu12/solidity-vuln-cls-integer-overflow-underflow-v1 --file contract.sol

Part of

This is one of 5 expert classifiers in the Solidity Vulnerability Detector system.

Expert	Hub Repo
Reentrancy	`jhsu12/solidity-vuln-cls-reentrancy-v1`
Access Control	`jhsu12/solidity-vuln-cls-access-control-v1`
Integer Overflow/Underflow	`jhsu12/solidity-vuln-cls-integer-overflow-underflow-v1`
Timestamp Dependence	`jhsu12/solidity-vuln-cls-timestamp-dependence-v1`
Unchecked Low-Level Calls	`jhsu12/solidity-vuln-cls-unchecked-low-level-calls-v1`