Unverified50% confidenceFactExact time

一个shape为(3,4)的连续张量，其stride为(4,1)

1

Sources

50%

Confidence

Long-term

Relevance

8/21/2026

First Seen

Sources

PyTorch Tensor Visualization: Understanding Tensor Operations Like Building with LEGO

Related Claims

Unverified形状为(3,1)的张量与形状为(1,4)的张量相加时，NumPy和PyTorch会自动广播为(3,4)的结果72% similar Unverified单张图片的shape为(3,224,224)时，模型期望的批次输入为(1,3,224,224)，需要unsqueeze(0)来添加batch维度71% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/781552

API

curl https://kongchang.com/api/v1/knowledge/claims/781552

MCP

get_claim(id=781552)