And so it is only existing in these quantized orbits.

只存在于这些量子化的轨道上。

But a piano is going to be quantized.

但是钢琴将会被量化。

And what that means is that it's quantize or comes in these small little quantities.

这意味着是量化或以这些小量出现。

In other words charge is quantized.

换句话说，电荷是量子化的。

For example, you replace things by discrete data and try to quantize it and so on.

例如，你将事物替换为离散数据，并尝试其进行量子化处理等。

And it sounds like a natural idea to quantize space.

将空间量子化听起来像是一个自然的想法。

And so electric charge is quantized.

电荷是量子化的。

A quantized amount of current into that loop.

进入该路的电流是一个量子化的量。

He also showed that photons are quantized.

他还表明光子是量子化的。

A flux sign is one of these quantized amounts of current that you can add to a loop.

一个通量符号是你可以在路添加的这些量子化的电流之一。

So for gravity, quantizing gravitational field, which is this curvature of space-time according to Einstein, you get another particle called graviton.

所以于引力， 将引力场量子化，根据爱斯坦的说法， 这就是时空的曲率，你会得到另一个被称为引力子的粒子。

And they also, if we look at them as a particle, the photoelectric effect tells us that photons are quantized.

而且， 如果我们把们视为粒子，光电效应告诉我们光子是量子化的。

And what I mean by quantized is it's going to come in discrete packets with a well-defined value of current.

我所说的“量子化”是指电流将以具有明确电流值的离散包形式流入。

The statistics of the images and then quantize them based on the statistics, much like you do in words, right?

图像的统计数据， 然后根据这些统计数据进行量化，就像你在处理文字时所做的那样， 吧？

So it's really important to be able to quantize or compress the model and be able to optimize memory that you use, you know, either during fine-tuning or training in LLM.

，能够量化或压缩模型并优化所使用的内存至关重要，无论是在大型语言模型（LLM）的微调还是训练过程。

Because the algorithm divides using these quantization tables, and then rounds to the nearest result, if we increase the values in the table, we will inevitably get more zeros in the resulting discrete cosine transformed and quantized block.

为该算法使用这些量化表进行划分，然后四舍五入到最接近的结果，如果我们增加表的值，我们不可避免地会在最终的离散余弦变换和量化块得到更多的零。