跟读练习: PyTorch in 100 Seconds - 通过视频学习英语口语
加载中...
1
PyTorch, an open-source deep learning framework used to build some of the world's most famous artificial intelligence products.
2
It was created at the Meta AI Research Lab in 2016, but is actually derived from the Lua-based Torch library that dates back to 2002.
3
Fundamentally, it's a library for programming with tensors, which are basically just multi-dimensional arrays that represent data and parameters in deep neural networks.
4
Sounds complicated, but its focus on usability will have you training machine learning models with just a few lines of Python.
5
In addition, it facilitates high-performance parallel computing on a GPU, thanks to NVIDIA's CUDA platform.
6
Developers love prototyping with it because it supports a dynamic computational graph, allowing models to be optimized at runtime.
7
It does this by constructing a directed acyclic graph consisting of functions
8
that keeps track of all the executed operations on the tensors, allowing you to change the shape, size, and operations after every iteration if needed.
9
PyTorch has been used to train models for computer vision AI like Tesla Autopilot, image generators like Stable Diffusion, and speech recognition models like OpenAI Whisper, just to name a few.
10
To get started, install PyTorch, and optionally CUDA, if you want to accelerate computing on your GPU.
11
Now import it into a Python file or notebook.
12
Like I mentioned, a tensor is similar to a multi-dimensional array.
13
Create a 2D array or matrix with Python, then use Torch to convert it into a tensor.
14
Now we can run all kinds of computations on it, like we might convert all these integers into random floating points.
15
We can also perform linear algebra by taking multiple tensors and multiplying them together.
16
What you came here to do though is build a deep neural network, like an image classifier.
17
To handle that, we can define a new class that inherits from the Neural Network Module class.
18
Inside the constructor, we can build it out layer by layer.
19
The flattened layer will take a multi-dimensional input, like an image, and convert it to one dimension.
20
From there, Sequential is used to create a container of layers that the data will flow through.
21
Each layer has multiple nodes, where each node is like its own mini statistical model.
22
As each data point flows through it, it'll try to guess the output, and gradually update a mapping of weights to determine the importance of a given variable.
23
Linear is a fully connected layer that takes the flattened 28x28 image and transforms it to an output of 512.
24
This layer is followed by a non-linear activation function.
25
When activated, it means that feature might be important and outputs the node.
26
Otherwise, it just outputs zero.
27
And finally, we finish with a fully connected layer that outputs the 10 labels the model is trying to predict.
28
With these pieces in place, the next step is to define a forward method that describes the flow of data.
29
And now instantiate the model to a GPU and pass it some input data.
30
This will automatically call its forward method for training and prediction.
31
Congratulations, you just built a neural network.
32
This has been PyTorch in 100 seconds.
33
Thanks for watching, and I will see you in the next one.
看视频学英语：通过技术讲解提升口语的实用场景
想象你正在观看一段关于人工智能框架PyTorch的英文视频，想要边学技术边练英语。这种结合专业内容的学习场景，既能积累行业词汇，又能通过真实语境提升听力和口语。视频中清晰的逻辑讲解、专业术语的使用，正是练习“shadowspeak”（影子跟读）的绝佳素材——你需要紧跟 speaker 的节奏，模仿发音、语调和停顿，在理解内容的同时强化口语输出。
必学实用表达与搭配
- open-source deep learning framework：开源深度学习框架（技术类文本高频搭配，注意“framework”的弱读）
- dynamic computational graph：动态计算图（专业术语，模仿时注意“computational”的重音位置）
- train machine learning models：训练机器学习模型（动词短语，“train”与“models”的连读要自然）
- non-linear activation function：非线性激活函数（科技英语中长定语的处理，模仿时保持语速稳定）
- forward method：前向传播方法（AI领域核心术语，“forward”的发音要准确）
你的影子跟读挑战：现在就行动！
打开视频，找到“Sounds complicated, but its focus on usability will have you training machine learning models with just a few lines of Python.”这段。先听3遍，注意 speaker 的语调变化（“but”后轻微升调，强调转折）和“usability”的发音。然后开始“shadow speech”训练：按下播放键后立即跟读，尽可能同步，重点模仿“just a few lines”的轻快节奏。重复5次后，录音对比自己与原音的差异，尤其关注“training”的重音和“Python”的尾音。完成后，你会发现不仅对技术内容更熟悉，英语发音和流畅度也有明显提升。记住，优质的“shadowing site”能提供多样素材，但关键是坚持——每天10分钟，用技术视频练口语，事半功倍！
什么是跟读法？
跟读法 (Shadowing) 是一种有科学依据的语言学习技巧，最初开发用于专业口译员的培训，并由多语言者Alexander Arguelles博士普及。这个方法简单而强大：您在听英语母语原声的同时立即大声重复——就像是一个延迟1-2秒紧跟说话者的影子。与被动听力或语法练习不同，跟读法强迫您的大脑和口腔肌肉同时处理并模仿真实的讲话模式。研究表明它能显着提高发音准确性，语调，节奏，连读，听力理解和口语流利度——使其成为雅思口语备考和真实英语交流最有效的方法之一。