零一万物Yi-1.5来了,国产开源大模型排行榜再次刷新

在 OpenAI 发布会的前一天,来自中国的大模型公司零一万物,让开源大模型社区活跃了起来:Yi 大模型家族新成员「Yi-1.5」模型正式开源。

在项目主页可以看到,Yi-1.5 包括一系列预训练和微调模型,分为 6B、9B、34B 三个版本,采用 Apache 2.0 许可证。

GitHub 地址:https://github.com/01-ai/Yi-1.5

Hugging Face 模型下载地址:https://huggingface.co/01-ai

据了解,Yi-1.5 是 Yi-1.0 的持续预训练版本,使用 500B 个 token 来提高编码、推理和指令执行能力,并在 300 万个指令调优样本上进行了精细调整。

与前序模型相比,Yi-1.5 系列模型进一步提升了编码、数学、推理和指令遵循能力。从下方多个基准测试结果可以看出,Yi-1.5 34B 型号的一些指标超过了 Qwen 的 72B,几乎与 Meta Llama 3 的 70B 相当。6B 和 9B 型号也成功超越了 Mistral 的 7B v0.2 版和 Gemma 的 7B 型号。

Yi-Large 进击全球SOTA大模型极品,Win Rate全球第一

在第三方权威评测中,零一万物 Yi 模型在全球头部大模型的中英文双语 PK 上表现出色。最新出炉的斯坦福评测机构 AlpacaEval 2.0 经官方认证的模型排行榜上,Yi-Large 模型的英语能力主要指标 LC Win Rate(控制回复的长度) 排到了世界第二,仅次于 GPT-4 Turbo,Win Rate 更排到了世界第一,此前国内模型中仅有 Yi 和 Qwen 曾经登上此榜单的前 20。





不过人家GPT4o目前都免费,Yi 却开启了收费。。。

  • 5
    点赞
  • 5
    收藏
    觉得还不错? 一键收藏
  • 0
    评论
To solve this problem, we can iterate over all possible center points of the protective field and calculate the minimum radius required to cover at least half of the creatures. Here's a step-by-step approach to solve the problem: 1. Read the input values: the number of creatures, n, and the coordinates of each creature. 2. Initialize variables to store the optimal position and radius of the field. Let's call them `optimal_x`, `optimal_y`, and `optimal_radius`. Set `optimal_radius` to a large value initially. 3. Iterate over all possible center points of the field. You can use nested loops to iterate over x and y coordinates within a reasonable range. 4. For each center point, calculate the distance from that point to all creatures. You can use the distance formula: `distance = sqrt((x - xi)^2 + (y - yi)^2)`. 5. Sort the distances in ascending order. 6. Calculate the radius required to cover at least half of the creatures. This can be done by taking the distance at index `ceil(n/2)` - 1, where `ceil()` is the ceiling function. 7. If the calculated radius is smaller than the current `optimal_radius`, update `optimal_x`, `optimal_y`, and `optimal_radius` with the current center point and radius. 8. After iterating over all possible center points, print the values of `optimal_x`, `optimal_y`, and `optimal_radius` as the output. Here's a Python implementation of the above approach: ```python import math n = int(input()) creatures = [] for _ in range(n): x, y = map(float, input().split()) creatures.append((x, y)) optimal_x = 0 optimal_y = 0 optimal_radius = float('inf') for x in range(-100, 101): for y in range(-100, 101): distances = [] for creature in creatures: xi, yi = creature distance = math.sqrt((x - xi)**2 + (y - yi)**2) distances.append(distance) distances.sort() radius = distances[math.ceil(n/2) - 1] if radius < optimal_radius: optimal_x = x optimal_y = y optimal_radius = radius print(optimal_x, optimal_y, optimal_radius) ``` This implementation has a time complexity of O(n^3) since it iterates over all possible center points and calculates the distances for each point. However, since the maximum value of n is 400, the solution should run efficiently within the given constraints.

“相关推荐”对你有帮助么?

  • 非常没帮助
  • 没帮助
  • 一般
  • 有帮助
  • 非常有帮助
提交
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值