8 added 1 character in body
source | link

I'm training a neural network but the training loss doesn't decrease. How can I fix this?

I'm not asking about overfitting or regularization. I'm asking about how to solve the problem where my network's performance doesn't improve on the training set.


This question is intentionally general so that other questions about how to train a neural network can be closed as a duplicate of this one, with the attitude that "if you give a man a fish you feed him for a day, but if you teach a man to fish, you can feed him for the rest of his life." See this metaMeta thread for a discussion: What's the best way to answer "my neural network doesn't work, please fix" questions?

If your neural network does not generalize well, see: What should I do when my neural network doesn't generalize well?

I'm training a neural network but the training loss doesn't decrease. How can I fix this?

I'm not asking about overfitting or regularization. I'm asking about how to solve the problem where my network's performance doesn't improve on the training set.


This question is intentionally general so that other questions about how to train a neural network can be closed as a duplicate of this one, with the attitude that "if you give a man a fish you feed him for a day, but if you teach a man to fish, you can feed him for the rest of his life." See this meta thread for a discussion: What's the best way to answer "my neural network doesn't work, please fix" questions?

If your neural network does not generalize well see: What should I do when my neural network doesn't generalize well?

I'm training a neural network but the training loss doesn't decrease. How can I fix this?

I'm not asking about overfitting or regularization. I'm asking about how to solve the problem where my network's performance doesn't improve on the training set.


This question is intentionally general so that other questions about how to train a neural network can be closed as a duplicate of this one, with the attitude that "if you give a man a fish you feed him for a day, but if you teach a man to fish, you can feed him for the rest of his life." See this Meta thread for a discussion: What's the best way to answer "my neural network doesn't work, please fix" questions?

If your neural network does not generalize well, see: What should I do when my neural network doesn't generalize well?

7 deleted 7 characters in body
source | link

I'm training a neural network but the training loss doesn't decrease. How can I fix this?

I'm not asking about overfitting or regularization. I'm asking about how to solve the problem where my network's performance doesn't improve on the training set.


This question is intentionally general so that other questions about how to train a neural network can be closed as a duplicate of this one, with the attitude that "if you give a man a fish you feed him for a day, but if you teach a man to fish, you can feed him for the rest of his life." See this meta thread for a discussion:

 What's the best way to answer "my neural network doesn't work, please fix" questions?

If your neural network does not generalize well see:

   What should I do when my neural network doesn't generalize well?

I'm training a neural network but the training loss doesn't decrease. How can I fix this?

I'm not asking about overfitting or regularization. I'm asking about how to solve the problem where my network's performance doesn't improve on the training set.


This question is intentionally general so that other questions about how to train a neural network can be closed as a duplicate of this one, with the attitude that "if you give a man a fish you feed him for a day, but if you teach a man to fish, you can feed him for the rest of his life." See this meta thread for a discussion:

 What's the best way to answer "my neural network doesn't work, please fix" questions?

If your neural network does not generalize well see:

 What should I do when my neural network doesn't generalize well?

I'm training a neural network but the training loss doesn't decrease. How can I fix this?

I'm not asking about overfitting or regularization. I'm asking about how to solve the problem where my network's performance doesn't improve on the training set.


This question is intentionally general so that other questions about how to train a neural network can be closed as a duplicate of this one, with the attitude that "if you give a man a fish you feed him for a day, but if you teach a man to fish, you can feed him for the rest of his life." See this meta thread for a discussion: What's the best way to answer "my neural network doesn't work, please fix" questions?

If your neural network does not generalize well see:  What should I do when my neural network doesn't generalize well?

6 added 185 characters in body
source | link

I'm training a neural network but the training loss doesn't decrease. How can I fix this?

I'm not asking about overfitting or regularization. I'm asking about how to solve the problem where my network's performance doesn't improve on the training settraining set.

 

(This This question is intentionally general so that other questions about how to train a neural network can be closed as a duplicate of this one, with the attitude that "if you give a man a fish you feed him for a day, but if you teach a man to fish, you can feed him for the rest of his life." See this meta thread for a discussion:  

What's the best way to answer "my neural network doesn't work, please fix" questions?)

If your neural network does not generalize well see:

What should I do when my neural network doesn't generalize well?

I'm training a neural network but the training loss doesn't decrease. How can I fix this?

I'm not asking about overfitting or regularization. I'm asking about how to solve the problem where my network's performance doesn't improve on the training set.

(This question is intentionally general so that other questions about how to train a neural network can be closed as a duplicate of this one, with the attitude that "if you give a man a fish you feed him for a day, but if you teach a man to fish, you can feed him for the rest of his life." See this meta thread for a discussion: What's the best way to answer "my neural network doesn't work, please fix" questions?)

I'm training a neural network but the training loss doesn't decrease. How can I fix this?

I'm not asking about overfitting or regularization. I'm asking about how to solve the problem where my network's performance doesn't improve on the training set.

 

This question is intentionally general so that other questions about how to train a neural network can be closed as a duplicate of this one, with the attitude that "if you give a man a fish you feed him for a day, but if you teach a man to fish, you can feed him for the rest of his life." See this meta thread for a discussion:  

What's the best way to answer "my neural network doesn't work, please fix" questions?

If your neural network does not generalize well see:

What should I do when my neural network doesn't generalize well?

    Notice removed Reward existing answer by amoeba
    Bounty Ended with Sycorax's answer chosen by amoeba
5 added 149 characters in body
source | link
    Notice added Reward existing answer by amoeba
    Bounty Started worth 50 reputation by amoeba
4 deleted 27 characters in body
source | link
    Tweeted twitter.com/StackStats/status/1008895757569744896
3 edited title
| link
2 added 313 characters in body
source | link
1
source | link