Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Why is my training loss so different from my test loss when i training wsj0-2mix #2303

Open
903859154 opened this issue Dec 15, 2023 · 3 comments
Open

Why is my training loss so different from my test loss when i training wsj0-2mix #2303

903859154 opened this issue Dec 15, 2023 · 3 comments
Labels
bug Something isn't working

Comments

@903859154
Copy link

Describe the bug

my trainning loss is 14 but my vaild and test loss is 8 . Why are the two so different? I saw your train loss is 22 and valid is 16. Is this wsj0-2mix' s bug?

Expected behaviour

I don't know why my loss is so low

To Reproduce

No response

Environment Details

No response

Relevant Log Output

No response

Additional Context

No response
@903859154 903859154 added the bug Something isn't working label Dec 15, 2023
@903859154
Copy link
Author

my network is conv-tasnet

@pplantinga
Copy link
Collaborator

Training loss most often exceeds validation/test loss due to data augmentation and regularization strategies such as dropout which make the task more difficult.

@pplantinga pplantinga reopened this May 30, 2024
@pplantinga
Copy link
Collaborator

On second read, are you asking why your loss is different from the official recipe?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
bug Something isn't working
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@pplantinga @903859154