In the case of Facebook, the company says data may hang around until the URL in question is reused, which is usually " after a short period of time." Though obviously that time can vary considerably.

在脸谱网的案例中，该公司表示相关数据可能以缓存的形式保留，直到URL地址被重新会显示出来，这种情况通常发生在发出删除命令后的短期内。但显然，公司在对这段时间的描述上与实际情况有很大的误差。

I used the computer before I went to the store, and when I came back, my history was gone, the cache was cleared, and the cookies had been reset.

离之前还在用电脑，等回来的时候，历史记录就没了，缓存被清空了，保存的密码也被重置了。

The defendant's Snipcam cache was always in the discovery, Your Honor.

被告的相册缓存一直是证据之一，法官大人。

The relay on this side cached them.

这边的存储器缓存了下来。

So we're making an nest box for the squirrel, and then we're going to also make him caching boxes as well.

所以们要给松鼠做一个巢箱，然后们也要给他做一个缓存箱。

It's got a 0.5 megabyte L2 backside cache, and then this thing screams, as you'll see in a minute.

它有一个0.5兆字节的L2后台缓存，你马上会听到存储器发出的声响。

It starts with a new microarchitecture that has a breakthrough feature we call Dynamic Caching, an industry first.

首先，它的新型微架构具备一个突破性的功能： 们首创的动态缓存。

And we've used our new architecture to improve parallelism and caching to reduce overall build times.

们还使用了新结构去增强并联和缓存，以此来缩短整体构建时间。

Cache spelled like this, but pronounced like the money, cash, okay?

缓存像这样拼写，​​但发音像钱，现金，好吗？

After the collection, the sample cache will be launched off the Red Planet.

采集完成后，样本缓存将被发射离这颗红色星球。

It's the success story of tourism, big contracts, and futuristic cache.

这是旅游业、大合同和未来缓存的成功故事。

It has them kind of cached in just on disk in Markdown files.

它将这些内容以Markdown文件的形式缓存在磁盘上。

This is slightly different from step caching.

这与步骤缓存略有不同。

And then some, a part of that is cached.

此外，其中一部分是缓存的。

We're going to have any kind of cat like level two cash?

们会达到任何类似二级缓存的层次吗？

Like before, instead of going back all the way to RAM to save that value, it can be stored in cached copy, which is faster to save to, and also faster to access later if more calculations are needed.

就像之前一样，不需要回到随机存取存储器去保存那个值，可以将其存储在缓存副本中，这样保存起来更快，如果后续需要更多的计算，访问也更快。

I know that the last so we run IPv6 internally for a lot of the routing from from our edge service, we are our architecture base as edge caches that sit in front of our origin servers for DNS.

知道，上次们内部为从边缘服务到DNS源服务器的许多路由都运行了IPv6。们的架构基础是位于DNS源服务器前端的边缘缓存。

If you want a dedicated deployment, there are also ways to, for example, keep the instances warm and cache the images on the node, or ways to make sure the model loads into memory faster.

如果你想进行专门的部署， 也有一些方法可以实现，比如保持实例的预热状态并在节点上缓存镜像，或者采取措施确保模型更快地加载到内存中。

You want to have some basic programming skills, but then you're going to dive into key concepts like reward modeling, supervised fine-tuning, mixture of experts layers, RMS, norm, rope, KV caching and more.

你需掌握一些基础编程技能，随后深入研究诸如奖励建模、监督微调、专家层混合、均方根归一化、旋转位置编码、KV缓存等关键概念。

The problem is: it's unlikely that a point would have started out as large as the points in these caches, that would be a lot of stone to chip away.

问题是：一个点始时不太可能像这些缓存中的点一样大，那将需要大量的石头。