Update Inference Engine checkpoint loading + meta tensor assertions #2940
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
lekurile
requested review from
RezaYazdaniAminabadi,
jeffra,
mrwyattii,
awan-10,
cmikeh2 and
arashb
as code owners
lekurile
changed the title
awan-10
reviewed
May 10, 2023
| @@ -151,6 +151,13 @@ def __init__(self, model, config): | |||
| assert pkg_version.parse(torch.__version__) >= pkg_version.parse("1.10"), \ | |||
| "If you want to use cuda graph, please upgrade torch to at least v1.10" | |||
|
|
|||
| # Check if model passed to engine is loaded w/ meta tensors, in which case | |||
| # kernel injection must be enabled. | |||
There was a problem hiding this comment.
Please update the comment to add why this only works for HF models.
There was a problem hiding this comment.
Added a note saying that the device type is sourced assuming a Hugging Face hierarchy.
awan-10
approved these changes
May 10, 2023
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Outdated suggestions cannot be applied.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.
This PR adds a
model_device_metaattribute to theInferenceEnginethat's used to:replace_with_kernel_inject == Trueif meta tensors are used, since we only support meta tensors when kernel injection is enabled.InferenceEngineto load checkpoints via_load_checkpoint()when a checkpoint is passed to theinit_inferenceAPI only when meta tensors are not used.This PR also adds an assertion in the
initialize_tensorsfunction of the base container to check that if the model is using meta tensors, that the corresponding model container uses the meta tensor feature.