Add eager flag and set compile to be always true #734
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
tgaddair
reviewed
Jan 9, 2025
| #[clap(long, env, value_enum)] | ||
| /// Only use either `--compile` or `--eager`. Using both at the same time will | ||
| /// result in an error. | ||
| #[clap(default_value = "true", long, env, value_enum)] |
There was a problem hiding this comment.
Since this defaults to true here, will this raise an error if the user specifies --eager without setting anything for compile?
Typically we would change the type of compile here to Option[bool] so we can distinguish between explicitly specified args vs unspecified.
| @@ -228,6 +228,10 @@ def check_initialized(self): | |||
| f"found uninitialized parameters in model {self.__class__.__name__}: {uninitialized_parameters}" | |||
| ) | |||
|
|
|||
| @property | |||
| def supports_cuda_graph_compilation(self) -> bool: | |||
| return True | |||
There was a problem hiding this comment.
Do we need to set this to False for some models?
There was a problem hiding this comment.
At least with the limited models I tested, compile worked so we don't set this property to False for any model, but we may want to in the future. Not needed at the moment so happy to get rid of it.
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Outdated suggestions cannot be applied.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.
Adds a new flag to the launcher called
--eagerthat allows users to disable CUDA graph compilation which is enabled by default. Also, sets--compileto always be true. Finally, tested that compile works with following list of models, and added a property calledsupports_cuda_graph_compilationthat individual models can set to disable compilation.