cassc

Fine-tuning a pre-trained model like bigcode/starencoder on a large collection of Solidity source code without any labeling can be done through unsupervised learning, specifically using masked language modeling (MLM). Here’s a step-by-step guide to fine-tuning the model for your specific needs:

Steps to Fine-Tune the Model

1. Prepare the Dataset

Ensure you have a large collection of Solidity source code files. Combine these files into a single or multiple text files.

Example: Combining Solidity Files into a Text File

cat *.sol > all_solidity_code.txt

cassc

Zero-shot examples refer to scenarios where a model is tasked with performing a specific task without having been explicitly trained on any data for that task. Instead, the model leverages its understanding from pre-training on other data to make predictions or inferences about the new task. This ability is particularly valuable for handling tasks or categories that were not seen during the model's training phase.

Zero-Shot Examples in Various Domains

1. Text Classification:

A model trained on general text data can classify text into categories it has never seen before by understanding the general concept of the categories.

• Example:
  • Input Text: "The weather today is sunny with a chance of showers in the evening."
• New Categories: "Weather Report" vs. "Financial News"

cassc

from typing import List
from pydantic import BaseModel, Field
from langchain_core.globals import set_debug
from langchain_core.output_parsers.json import JsonOutputParser
from langchain_core.prompts.prompt import PromptTemplate
from langchain_openai import ChatOpenAI
from langchain_community.callbacks import get_openai_callback
from langchain.chains import ConversationChain
from langchain_community.llms import OpenAI

cassc

from langchain_core.globals import set_debug
from langchain_openai import ChatOpenAI
from langchain_community.callbacks import get_openai_callback
from langchain.chains import ConversationChain
from langchain.chains import ConversationChain
from langchain.memory import ConversationBufferMemory, ConversationSummaryBufferMemory


set_debug(False)
model_name = 'gpt-4o'

cassc

tokio vs threadpool

Tokio is an asynchronous runtime for the Rust programming language, and it is more than just thread pools. Here's a brief overview of how Tokio is implemented:

1. Event Loop: Tokio uses an event-driven architecture, leveraging an event loop to manage tasks and I/O events. The core of Tokio's runtime is the mio library, which provides a low-level event loop backed by OS-specific mechanisms like epoll (Linux), kqueue (BSD), and IOCP (Windows).

2. Futures and Tasks: Tokio's concurrency model is based on Rust's Future trait. Tasks are units of work that implement the Future trait, and they are polled by the runtime to make progress. When a task is not ready to make progress (e.g., waiting for I/O), it yields control back to the runtime.

3. Thread Pool:

cassc

{
  "00000014-mixed-4": {
    "env": {
      "currentCoinbase": "b94f5374fce5edbc8e2a8697c15331677e6ebf0b",
      "currentDifficulty": "0x200000",
      "currentRandom": "0x0000000000000000000000000000000000000000000000000000000000200000",
      "currentGasLimit": "0x26e1f476fe1e22",
      "currentNumber": "0x1",
      "currentTimestamp": "0x3e8",
      "previousHash": "0x044852b2a670ade5407e78fb2863c51de9fcb96542a07186fe3aeda6bb8a116d",

cassc

{
  "00000007-mixed-4": {
    "env": {
      "currentCoinbase": "b94f5374fce5edbc8e2a8697c15331677e6ebf0b",
      "currentDifficulty": "0x200000",
      "currentRandom": "0x0000000000000000000000000000000000000000000000000000000000200000",
      "currentGasLimit": "0x26e1f476fe1e22",
      "currentNumber": "0x1",
      "currentTimestamp": "0x3e8",
      "previousHash": "0x044852b2a670ade5407e78fb2863c51de9fcb96542a07186fe3aeda6bb8a116d",

cassc

{
    "00000001-mixed-9": {
        "env": {
            "currentCoinbase": "b94f5374fce5edbc8e2a8697c15331677e6ebf0b",
            "currentDifficulty": "0x200000",
            "currentRandom": "0x0000000000000000000000000000000000000000000000000000000000200000",
            "currentGasLimit": "0x26e1f476fe1e22",
            "currentNumber": "0x1",
            "currentTimestamp": "0x3e8",
            "previousHash": "0x044852b2a670ade5407e78fb2863c51de9fcb96542a07186fe3aeda6bb8a116d",

cassc

// cuevm --input xxx.json
// geth --json --noreturndata --nomemory statetest xxx.json

{
    "00000004-mixed-1": {
        "env": {
            "currentCoinbase": "b94f5374fce5edbc8e2a8697c15331677e6ebf0b",
            "currentDifficulty": "0x200000",
            "currentRandom": "0x0000000000000000000000000000000000000000000000000000000000200000",
            "currentGasLimit": "0x26e1f476fe1e22",

cassc

# ❯ CAST_BIN=[optional-cast-binary-full-path] python cast-json-example.py

import subprocess
import json
import os
import tempfile

cast_bin = os.environ.get('CAST_BIN', 'cast')

def cast_run(rpc_url, txhash):