12
votes
Accepted
What is the relationship between gradient accumulation and batch size?
There isn't any explicit relation between the batch size and the gradient accumulation steps, except for the fact that gradient accumulation helps one to fit models with relatively larger batch sizes (...
11
votes
Accepted
Effect of batch size and number of GPUs on model accuracy
This should make a difference, but how big is the difference heavily depends on your task. However, generally speaking, a smaller batch size will have a lower speed if counted in sample/minutes, but ...
7
votes
How do I choose the optimal batch size?
From the blog A Gentle Introduction to Mini-Batch Gradient Descent and How to Configure Batch Size (2017) by Jason Brownlee.
How to Configure Mini-Batch Gradient Descent
Mini-batch gradient descent ...
Community wiki
7
votes
How do I choose the optimal batch size?
Here are a few guidelines, inspired by the deep learning specialization course, to choose the size of the mini-batch:
If you have a small training set, use batch gradient descent (m < 200)
In ...
5
votes
Accepted
Isn't it useless to batch with repetition in RL?
Pythons random.choice may return the same sample within a single batch. Although you could view this as redundant data, it is still a fair sample statistically, so ...
5
votes
What is the purpose of the batch size in neural networks?
tl;dr: A batch size is the number of samples a network sees before updating its gradients. This number can range from a single sample to the whole training set. Empirically, there is a sweet spot in ...
4
votes
Is it true that batch size of form $2^k$ gives better results?
The choice of the batch size to be a power of 2 is not due the quality of predictions .
The larger the batch_size is - the better is the estimate of the gradient, but a noise can be beneficial to ...
4
votes
How do I choose the optimal batch size?
The batch size can also have a significant impact on your model’s performance and the training time. In general, the optimal batch size will be lower than 32. In April 2018, Yann Lecun even tweeted:
...
3
votes
Is there a way to translate the concept of batch size into reinforcement learning?
Potentially.
If you do offline reinforcement learning, you're basically learning to approximate a function by sampling input/output pairs, rather than episode-by-episode. Here, your batch size could ...
3
votes
Effect of batch size and number of GPUs on model accuracy
No. Different batch sizes mean different gradients (check stochastic gradient descent concept you will get how loss calculated) are calculated in each step, and thus the gradient descent will likely ...
3
votes
Accepted
What is the purpose of a replay/memory buffer in Deep Q-Learning networks?
This question is asking about randomness in the data, so I will first clarify one of the comments in this question made about randomness.
If the data is taken randomly, how can a DQN approximate a ...
2
votes
Are there any rules for choosing batch size?
Data that has a size of a factor of $2$ (aka $2^n$ for some integer $n$) allows for easier memory management because data can be organized contiguously (without gaps). This allows for faster memory ...
2
votes
Accepted
What is the difference between validation percentage and batch size?
The percentages refer to the number of samples to use (out of full dataset) as the validation and test datasets. So if you pass a dataset that consists of 100,000 samples to the model and set the ...
2
votes
Accepted
What is the reason behind using a test batch size?
I am not familiar with using batches during network evaluation. Can someone explain what is the reason behind using it and what are advantages and disadvantages?
It is usually just for memory use ...
2
votes
Is there a way to translate the concept of batch size into reinforcement learning?
From my understanding of reinforcement learning, you will have an agent and an environment.
In each episode, the agent observes the state $s$, takes some action action $a$, then gets some reward $r$, ...
2
votes
Is there any relationship between the batch size and the number of epochs?
The smaller the batch_size is, the larger number of batches is processed per one epoch.
On one hand, since one makes more steps per epoch, one can think, that less ...
2
votes
Why does linearly decreasing batch sizes result in exponentially increasing training times?
This has a very simple hardware explanation.
GPUs have several thousand cores. If you are not making use of enough of these cores, eventually the cost of moving data over to the gpu comes to dominate ...
1
vote
Training a neural network in full batch training
I think the closest approach to what is described in the papers you linked is Neural Style Transfer. But I see also some (potential) misunderstanding regarding the full batch training, so let me ...
1
vote
Why should data batches in a neural network have an equal size?
They don't have to be equal, but introducing variable batches brings more complexity.
In particular you need to make sure that batches are representative enough of the full dataset. If the batches ...
1
vote
What can be an example other than batch normalization that uses statistics of batches?
Here's some examples:
Group Normalization
Layer Normalization
Switchable Normalization
Attentive Normalization
Spectrl Normalization
Notice how in general different normalization techniques are ...
1
vote
Is it true that batch size of form $2^k$ gives better results?
The main reason to use powers of 2 is in the way existing hardware and software are made, there isn't any purely mathematical reason. CPUs, GPUs, memories, and internal buses all use a size that's the ...
1
vote
Will there be any changes in the model's performance due to the usage of very small batch sizes?
Batch size affects how many training updates (steps) will happen during each epoch.
When the batch size is small, this means that the model sees fewer data in each weights update. Thus, your question ...
Only top scored, non community-wiki answers of a minimum length are eligible
Related Tags
batch-size × 21neural-networks × 6
deep-learning × 6
training × 3
hyper-parameters × 3
machine-learning × 2
reinforcement-learning × 2
convolutional-neural-networks × 2
comparison × 2
transformer × 2
stochastic-gradient-descent × 2
gpu × 2
tensorflow × 1
keras × 1
recurrent-neural-networks × 1
dqn × 1
backpropagation × 1
gradient-descent × 1
deep-neural-networks × 1
attention × 1
variational-autoencoder × 1
performance × 1
batch-normalization × 1
accuracy × 1
cross-validation × 1