README.md

metadata

language:
  - en
license: apache-2.0
tags:
  - solidity
  - web3
  - code generation
widget:
  - text: |-
      pragma solidity ^0.5.7;
      // Context: ParentA | Functions: helloA helloB | Constants: constantA 
      contract HelloWorld is ParentA {

A code autocomplete T5 model for solidity

• Hello world example to use this model, notice the input text includes
  • Header solidity version like pragma solidity ^0.5.7
  • Ancestor class/library info, e.g. public functions and constants from ParentA
  • Contract/Library/Interface declaration header, e.g. HelloWorld ended with {

from transformers import AutoTokenizer, T5ForConditionalGeneration
tokenizer = AutoTokenizer.from_pretrained("hululuzhu/solidity-autocomplete")
model = T5ForConditionalGeneration.from_pretrained("hululuzhu/solidity-autocomplete")

text = """pragma solidity ^0.5.7;
// Context: ParentA | Functions: helloA helloB | Constants: constantA 
contract HelloWorld is ParentA {"""
input_ids = model.tokenizer(text, return_tensors="pt", truncation=True).input_ids.to('cuda')

# Need to tune beam/topk/topp params to get good outcome
generated_ids = model.model.generate(input_ids, max_length=256, num_beams=5, top_p=0.95, top_k=50)
print(model.tokenizer.decode(generated_ids[0], skip_special_tokens=True))

• Base T5 code model: https://huggingface.co/Salesforce/codet5-large
• Source data: https://huggingface.co/datasets/mwritescode/slither-audited-smart-contracts

  • Processing steps: Clean, contract-level segmentation sepration, split in and out

  • After processing input sample

    pragma solidity 0.5.7;
    // Context: PauserRole | Functions: isPauser addPauser renouncePauser | Constants: 
    contract Pausable is PauserRole {
    

  • After processing output sample (notice indentation is bad, this is intentional to reduce token size)

    event Paused(address account);
    event Unpaused(address account);
    bool private _pausableActive;
    bool private _paused;
    constructor () internal {
    _paused = false;
    }
    function paused() public view returns (bool) {
    return _paused;
    }
    modifier whenNotPaused() {
    require(!_paused);
    _;
    }
    modifier whenPaused() {
    require(_paused);
    _;
    }
    function pause() public onlyPauser whenNotPaused whenPausableActive {
    _paused = true;
    emit Paused(msg.sender);
    }
    function unpause() public onlyPauser whenPaused whenPausableActive {
    _paused = false;
    emit Unpaused(msg.sender);
    }
    function _setPausableActive(bool _active) internal {
    _pausableActive = _active;
    }
    modifier whenPausableActive() {
    require(_pausableActive);
    _;
    }
    }
    

• Source training code: To be added