指令 / SFT · BAAI

Infinity-Instruct

A 7.5M high-quality instruction dataset from the Beijing Academy of AI, combining foundational skills (math, code, reasoning) with conversational ability. Used to train models that surpass Llama 3 70B while being much smaller.

加载
from datasets import load_dataset
ds = load_dataset("BAAI/Infinity-Instruct")
预览示例数据
{
  "id": "inf_001",
  "conversations": [
    {"from": "human", "value": "Write a Python function to find all prime numbers up to n using the Sieve of Eratosthenes."},
    {"from": "gpt", "value": "def sieve_of_eratosthenes(n):\n    primes = [True] * (n+1)\n    primes[0] = primes[1] = False\n    for i in range(2, int(n**0.5)+1):\n        if primes[i]:\n            for j in range(i*i, n+1, i):\n                primes[j] = False\n    return [i for i in range(2, n+1) if primes[i]]"}
  ]
}

用这个数据集微调

使用 QLoRA(4-bit 基础模型 + LoRA 适配器)微调的预计显存需求(保守默认参数):

7B QLoRA · ~6GB VRAM13B QLoRA · ~10GB VRAM
检测你的 GPU 能否微调 →

微调新手?跟着分步教程走: 一小时微调你的第一个 LLM

常见问题

Infinity-Instruct 可以商用吗?

可以——Infinity-Instruct 采用 Apache 2.0 宽松许可证,允许商业使用,包括训练用于产品的模型。发布前请查看数据集卡片中的署名要求。

Infinity-Instruct 有多少数据?需要全部使用吗?

Infinity-Instruct 包含 7.5M Instructions。通常不需要全部:风格和格式微调只需几百到几千条样本——先加载切片(如 split="train[:1000]"),质量到达瓶颈时再扩大规模。

Infinity-Instruct 最适合做什么?

Large-scale general SFT when you need millions of samples。它属于数据集中心的「指令 / SFT」板块,那里有替代和互补的数据集。