Skip to content

Instantly share code, notes, and snippets.

@ozgureyilmaz
ozgureyilmaz / grpo_demo.py
Created January 30, 2025 22:09 — forked from willccbb/grpo_demo.py
GRPO Llama-1B
# train_grpo.py
import re
import torch
from datasets import load_dataset, Dataset
from transformers import AutoTokenizer, AutoModelForCausalLM
from peft import LoraConfig
from trl import GRPOConfig, GRPOTrainer
# Load and prep dataset
// SPDX-License-Identifier: MIT
pragma solidity ^0.8.0;
// Chain adapter interface
interface IChainAdapter {
function getChainId() external view returns (uint256);
function getNativeToken() external view returns (address);
function getBridgeAddress() external view returns (address);
function getFeeToken() external view returns (address);
function getGasPrice() external view returns (uint256);
Instruction Name RV32I RV64I RV128I Syntax
Loads
LB Load Byte Y Y Y LB rd,rs1,imm
LH Load Halfword Y Y Y LH rd,rs1,imm
LW Load Word Y Y Y LW rd,rs1,imm
LD Load Double N Y Y LD rd,rs,imm
LQ Load Quad N N Y LQ rd,rs,imm
LBU Load Byte Unsigned Y Y Y LBU rd,rs1,imm
LHU Load Halfword Unsigned Y Y Y LHU rd,rs1,imm
Register ABI Name Description
x0 zero zero register
x1 ra return address
x2 sp stack pointer
x3 gp global pointer
x4 tp thread pointer
x5 t0 temporary register
x6 t1 temporary register
x7 t2 temporary register
// SPDX-License-Identifier: MIT
pragma solidity ^0.8.0;
contract HashExample {
function getHash(bytes memory data) public pure returns (bytes32) {
return sha256(data);
}
}
# Mnemonic | Instruction | Action |
ADC Add with carry Rd := Rn + Op2 + Carry
ADD Add Rd := Rn + Op2
AND AND Rd := Rn AND Op2
B Branch R15 := address
BIC Bit Clear Rd := Rn AND NOT Op2
BL Branch with Link R14 := R15, R15 := address
BX Branch and Exchange R15 := Rn, T bit := Rn[0]
CDP Coprocesor Data Processing (Coprocessor-specific)