# 🚚 EMA#

Exponential Moving Average (EMA) is a model averaging technique that maintains an exponentially weighted moving average of the model parameters during training. The averaged parameters are used for model evaluation. EMA typically results in less noisy validation metrics over the course of training, and sometimes increased generalization.

## How to Use#

### Functional Interface#

# Run the EMA algorithm directly on the batch data using the Composer functional API

import copy

import composer.functional as cf

loss_fn = F.cross_entropy
ema_model = copy.deepcopy(model)
model.train()

for epoch in range(num_epochs):
y_hat = model(X)
loss = loss_fn(y_hat, y)
loss.backward()
opt.step()
cf.compute_ema(model, ema_model, smoothing=0.99)


### Composer Trainer#

# Instantiate the algorithm and pass it into the Trainer
# The trainer will automatically run it at the appropriate points in the training loop

from composer.algorithms import EMA
from composer.trainer import Trainer

ema = EMA(half_life='50ba')

trainer = Trainer(model=model,
max_duration='1ep',
algorithms=[ema])

trainer.fit()

model = ema.ema_model


### Implementation Details#

Because EMA needs to maintain a copy of the model’s (averaged) weights, it requires a bit more on-device memory. In the functional implementation, the amount of extra memory is 2x the size of the model. In the composer trainer implementation, it is 3x the size of the model to allow for swapping the training and evaluation models. In practice, the extra memory used is small relative to the total amount of memory used, as activations and optimizer state are not duplicated.

EMA also uses a bit of extra compute to calculate the moving average. This can lead to a small slowdown. The extra compute can be reduced by not computing the moving average every iteration. In the composer trainer implementation this can be done by using a larger update_interval. In practice we find that as long as half_life is much larger than update_interval, increasing update_interval does not have much effect on generalization performance.

## Suggested Hyperparameters#

The Composer Trainer implementation of EMA has several hyperparameters:

• half_life - The half life for terms in the average. A longer half life means old information is remembered longer, a shorter half life means old information is discared sooner. Defaults to '1000ba'

• update_interval - The period at which updates to the moving average are computed. A longer update interval means that updates are computed less frequently. If left unspecified, this defaults to 1 in the units of half_life, or 1ba if using smoothing.

• ema_start - The amount of training completed before SWA is applied. The default value is '0.0dur' which starts EMA at the start of training.

A good typical starting value for half_life is half_life="1000ba", for a half life of 1000 batches. At the same time, update_interval can be left unspecified which will default to update_interval="1ba", or set to a larger value such as update_interval="10ba" to improve runtime. Shorter update intervals typically result in better generalization performance at the cost of somewhat increased runtime.

For compatibility with other implementations, there is also an option to specify the value of smoothing directly.

• smoothing - The coefficient representing the degree to which older observations are kept. The default (unspecified) value is None. Should only be used if half_life is not used

To use this, half_life should be set to half_life=None, and the value of smoothing given instead. This value is not modified when update_interval is changed, and so changes to update_interval when using smoothing will result in changes to the time scale of the average.

## Technical Details#

✅ EMA Improves the Tradeoff Between Quality and Training Speed

In our experiments, EMA improves the attainable tradeoffs between training speed and the final quality of the trained model. We recommend EMA for training convolutional networks.

✅ EMA should result in less noisy validation metrics during training

If evalutation metrics are computed over the course of training, EMA should result in these metrics being smoother and less noisy due to averaging.

🚧 Composing Model-Averaging Methods

As a general rule, model-averaging methods do not compose well. We recommend using one of EMA or SWA, but not both.

❗ EMA increases memory consumption

Because EMA needs to maintain a copy of the model’s (averaged) weights, it requires a bit more on device memory. In practice, the extra memory used is small relative to the total amount of memory used, as activations and optimizer state are not duplicated.

❗ EMA uses some extra compute

This can lead to a small slowdown. The extra compute can be reduced by not computing the moving average every iteration. In the composer trainer implementation this can be done by using a larger update_interval.

❗ Evaluation should not be done with the training model

Evaluation should be done with the ema_model in the functional impementation as this is the model containing the averaged parameters. The ema model can be accessed after training from the EMA object via model = ema.ema_model in the composer trainer implementation. Similarly, the model without ema applied (the training model) can be accessed via model=ema.training_model. By default, when saving checkpoints with the CheckpointSaver callback or through trainer arguments the weights saved will be the ema model weights. An exception is if saving is done by explicitly calling trainer.save_checkpoint() which will result in the training model weights being saved as state.model.

Algorithm class: composer.algorithms.EMA
Functional: composer.functional.compute_ema()