What is the significance of "p" in TripletLoss

[ani0135]

System information (version)

• Pytorch => :grey_question:
• Operating System / Platform => :grey_question:
• Cuda => :grey_question:

Detailed description

What is the significance of "p" in TripletLoss? In forward of TripletLoss embeddings: [n, c, p], label: [n] and this is change into # [n, c, p] -> [p, n, c], why this permute has been done for embeddings and what is significance of "p" here?

Steps to reproduce

Issue submission checklist

• [x] I checked the problem with documentation, FAQ, issues, and have not found solution.

[ChaoFan996]

Hi, here 'p' means the part number. This permutation means that we will treat each part in parallel, i.e. separately.