diffusion_policy
diffusion_policy copied to clipboard
Is EMA used to update the training weights?
Hello @cheng-chi ,
In your colab notebook I find that you mention the EMA model can "accelerate training and improves stability":
However, in the training loop, you may only update the EMA model with the training model:
The ema.step function will not update/smooth the training weights. Why can it accelerate training?
If my understanding is wrong, please let me know
Best, StarCycle