[autoparallel] Add metainfo support for F.linear #1987
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Merge ColossalAI
Daily merge
…r30/ColossalAI into feature/metainfo_for_auto_parallel
…r30/ColossalAI into feature/metainfo_for_auto_parallel
Cypher30
commented
Nov 18, 2022
| @@ -65,7 +68,7 @@ def linear_meta_info(*args, **kwargs) -> Tuple[TrainCycleItem, TrainCycleItem, L | |||
| has_bias: bool = False | |||
| input_tensor = next(filter(lambda x: x.type == OperationDataType.ARG, args)).data | |||
| output_tensor = next(filter(lambda x: x.type == OperationDataType.OUTPUT, args)).data | |||
| weight_tensor = next(filter(lambda x: x.name == 'weight', args)).data | |||
| weight_tensors = [x.data for x in args if x.type == OperationDataType.PARAM] | |||
There was a problem hiding this comment.
Modify this part for more robust code
Cypher30
commented
Nov 18, 2022
| assert meta_register.has(self._target.__class__), f'{self._target.__class__} not found in the meta registry' | ||
| meta_func = meta_register.get(self._target.__class__) | ||
| try: | ||
| # module |
There was a problem hiding this comment.
Support the case that node.op == “call_function”
Cypher30
commented
Nov 18, 2022
| @@ -104,8 +106,12 @@ def mem_test_for_node_strategy(rank: int, | |||
| ) | |||
|
|
|||
| # estimated memory | |||
| metainfo = MetaInfo(target_node.strategies_vector[strategy_index], | |||
| target_node.graph.owning_module.get_submodule(target_node.target)) | |||
| if target_node.op == "call_module": | |||
There was a problem hiding this comment.
Modify this part to support node.op == “call_function”
YuliangLiu0306
approved these changes
Nov 23, 2022
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
What’s New?
In this PR, I done some work to support
torch.nn.functional.linearin our metainfo generation, the memory estimation results are aligned withtorch.nn.Linear(without bias). And though the biased linear is now separated by us into matmul and bias add, I still retain the part that generate metainfo for biased linear for future use.