A recipe for training neural networks

http://karpathy.github.io/2019/04/25/recipe/

Some few weeks ago I posted a tweet on “the most common neural net mistakes”, listing a few common gotchas related to training neural nets. The tweet got quite a bit more engagement than I anticipated (including a webinar :)). Clearly, a lot of people have personally encountered the large gap between “here is how a convolutional layer works” and “our convnet achieves state of the art results”.

I wanted to dig a bit deeper and talk about how one can avoid making these errors altogether (or fix them very fast). The trick to doing so is to follow a certain process, which as far as I can tell is not very often documented. Let’s start with two important observations that motivate it.

1) Neural net training is a leaky abstraction

2) Neural net training fails silently

Everything could be correct syntactically, but the whole thing isn’t arranged properly, and it’s really hard to tell. The “possible error surface” is large, logical (as opposed to syntactic), and very tricky to unit test. For example, perhaps you forgot to flip your labels when you left-right flipped the image during data augmentation. Your net can still (shockingly) work pretty well because your network can internally learn to detect flipped images and then it left-right flips its predictions.

The recipe

In light of the above two facts, I have developed a specific process for myself that I follow when applying a neural net to a new problem, which I will try to describe. You will see that it takes the two principles above very seriously. In particular, it builds from simple to complex and at every step of the way we make concrete hypotheses about what will happen and then either validate them with an experiment or investigate until we find some issue. What we try to prevent very hard is the introduction of a lot of “unverified” complexity at once, which is bound to introduce bugs/misconfigurations that will take forever to find (if ever).

1. Become one with the data

The first step to training a neural net is to not touch any neural net code at all and instead begin by thoroughly inspecting your data. This step is critical. I like to spend copious amount of time (measured in units of hours) scanning through thousands of examples, understanding their distribution and looking for patterns. Luckily, your brain is pretty good at this.

2. Set up the end-to-end training/evaluation skeleton + get dumb baselines

Our next step is to set up a full training + evaluation skeleton and gain trust in its correctness via a series of experiments. At this stage it is best to pick some simple model that you couldn’t possibly have screwed up somehow - e.g. a linear classifier, or a very tiny ConvNet. We’ll want to train it, visualize the losses, any other metrics (e.g. accuracy), model predictions, and perform a series of ablation experiments with explicit hypotheses along the way.

• fix random seed; simplify; verify loss @ init; init well; overfit one batch; visualize just before the net; visualize prediction dynamics …

3. Overfit

At this stage we should have a good understanding of the dataset and we have the full training + evaluation pipeline working. For any given model we can (reproducibly) compute a metric that we trust. We are also armed with our performance for an input-independent baseline, the performance of a few dumb baselines (we better beat these), and we have a rough sense of the performance of a human (we hope to reach this). The stage is now set for iterating on a good model.

The approach I like to take to finding a good model has two stages: first get a model large enough that it can overfit (i.e. focus on training loss) and then regularize it appropriately (give up some training loss to improve the validation loss). The reason I like these two stages is that if we are not able to reach a low error rate with any model at all that may again indicate some issues, bugs, or misconfiguration.

• adam is safe; complexify only one at a time; do not trust learning rate decay defaults …

4. Regularize

Ideally, we are now at a place where we have a large model that is fitting at least the training set. Now it is time to regularize it and gain some validation accuracy by giving up some of the training accuracy. Some tips & tricks:

• get more data; data augment; creative augmentation; pretrain; stick with supervised learning; smaller input dimensionality; smaller model size; decrease the batch size; drop; weight decay …

5. Tune

You should now be “in the loop” with your dataset exploring a wide model space for architectures that achieve low validation loss. A few tips and tricks for this step:

• random over grid search; hyper-parameter optimization

6. Squeeze out the juice

Once you find the best types of architectures and hyper-parameters you can still use a few more tricks to squeeze out the last pieces of juice out of the system:

• ensembles; leave it training