Skip to content

Instantly share code, notes, and snippets.

@BakiBaltas
BakiBaltas / grpo_demo.py
Created January 31, 2025 05:24 — 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
$prefixes: ("-webkit-","-moz-", "-o-", "-ms-", "");
@mixin selection($color, $background, $textshadow: none) {
@each $prefix in $prefixes {
::#{$prefix}selection {
color: $color;
background: $background;
text-shadow: none;
}
}