导出为 ONNX
让模型在不同运行时中运行
导出为 ONNX 是 CoddyKit 上的免费 Deep Learning Academy 课时。 这是第 2 节课，共 4 节。 你可以在下方免费阅读本课时的完整内容 — 然后在浏览器中使用内置代码编辑器和全天候 AI 导师进行实践。 这是 Deep Learning Academy 学习路径的一部分，你的进度在网页和 CoddyKit 应用中同步。 Deep Learning Academy 课程共包含 4 节课。
本课时的部分内容尚未翻译，以英文显示。
One Model, Many Runtimes
Sometimes your model must run somewhere PyTorch is not installed. ONNX is a shared format that lets one model run across many engines. 🌐
What ONNX Stands For
ONNX means Open Neural Network Exchange. It is a vendor-neutral file format that describes your model as a graph of standard operations.
Why Teams Love ONNX
With ONNX you train in PyTorch but deploy on ONNX Runtime, mobile, or the browser, without rewriting the model for each platform.
Export in One Call
You export by giving PyTorch the model and a sample input. The exporter traces the run and writes a portable graph to disk.
import torch
torch.onnx.export(model, sample_input, 'model.onnx')The Dummy Input Shapes the Graph
That sample tensor must match your real input shape and dtype, because the exporter records the operations the data triggers.
sample_input = torch.randn(1, 3, 224, 224)Name Your Inputs and Outputs
Give clear input and output names so the serving code can bind data by name instead of guessing positions.
torch.onnx.export(model, x, 'm.onnx',
input_names=['image'], output_names=['logits'])Allow Flexible Batch Sizes
By default the batch size is fixed. Mark it as a dynamic axis so the exported model accepts any number of inputs at once.
dynamic_axes={'image': {0: 'batch'}}Always Check Your Export
Load the file with the onnx library and run the built-in checker to confirm the graph is valid before you ship it. ✅
import onnx
onnx.checker.check_model(onnx.load('model.onnx'))Run It with ONNX Runtime
ONNX Runtime is a fast engine that executes the exported model on CPU or GPU, often quicker than plain Python inference.
import onnxruntime as ort
session = ort.InferenceSession('model.onnx')Confirm the Numbers Match
Run the same input through PyTorch and ONNX Runtime and compare outputs. They should match closely, proving the export is faithful.
Mind the Opset Version
Each export targets an opset version, the set of supported operations. Pick a version your target runtime understands to avoid errors.
torch.onnx.export(model, x, 'm.onnx', opset_version=17)Quick Check
You want a fixed batch to instead accept any size. What do you set?
Recap: Portable Across Runtimes
You exported a PyTorch model to ONNX with a sample input, named axes, validated it, and ran it on ONNX Runtime anywhere. 🎉
常见问题解答
「导出为 ONNX」课时是免费的吗？
是的 — 「导出为 ONNX」的完整文本可在网页上免费阅读。要进行交互式练习（内置代码编辑器和全天候 AI 导师）并解锁 Deep Learning Academy 课程的其余内容，请升级到 CoddyKit PRO。 Deep Learning Academy 课程共包含 4 节课。
「导出为 ONNX」这节课中我会学到什么？
让模型在不同运行时中运行 你通过在浏览器中直接运行的动手代码来练习 Deep Learning Academy，全天候 AI 导师会在你学习这节课的过程中回答你的问题。
学习 Deep Learning Academy 需要有经验吗？
无需任何先前经验。CoddyKit 上的 Deep Learning Academy 课程适合初学者到高级学习者，你可以从这里开始或从头开始，按照自己的节奏学习。 这是第 2 节课，共 4 节。
「导出为 ONNX」课时需要多长时间？
大多数 CoddyKit 课程大约需要 5–10 分钟。每节课都很精短且互动，所以你能稳步进步，并在网页和应用中从离开的地方继续。
我能在这节 Deep Learning Academy 课中编写并运行代码吗？
能。每节 Deep Learning Academy 课都包含内置代码编辑器，你可以在浏览器中直接编写并运行真实代码，并获得即时 AI 反馈 — 无需本地设置。