Support device_map=sequential & max_memory config parameters #903
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
In order to train a large model, I recently had to use the combination of
device_map=sequential
andmax_memory
in order to balance the memory usage per gpu (in model parallel) manually.device_map=auto
should not be forced when a value has been specified in the configmax_memory
allows for setting memory limits per gpu and is most useful withdevice_map=sequential
This was the simplest way I found to implement supporting these two options.
cc @winglian