[python] Rolling batch support for flash models #865
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
engines/python/setup/djl_python/rolling_batch/lmi_dist_rolling_batch.py
Outdated
Show resolved
Hide resolved
xyang16
changed the title
xyang16
changed the title
lanking520
reviewed
Jun 28, 2023
| self.rolling_batch = SchedulerRollingBatch(model_id_or_path, | ||
| self.device, properties, | ||
| **kwargs) | ||
| if os.getenv('OMPI_COMM_WORLD_SIZE'): |
There was a problem hiding this comment.
check for engine, if engine != Python and tensor_parallel!=1
lanking520
reviewed
Jun 28, 2023
| if self.enable_rolling_batch and self.enable_rolling_batch.lower( | ||
| ) == "false": | ||
| self.enable_rolling_batch = None | ||
| self.enable_rolling_batch = properties.get("rolling_batch") is not None |
There was a problem hiding this comment.
self.enable_rolling_batch = properties.get("rolling_batch", None)
| **kwargs) | ||
| if self.engine != "Python" and tp_degree != 1: | ||
| self.device = int(os.getenv("LOCAL_RANK", 0)) | ||
| rolling_batch_type = properties.get("rolling_batch") |
| self.device = int(os.getenv("LOCAL_RANK", 0)) | ||
| rolling_batch_type = properties.get("rolling_batch") | ||
| model_config = AutoConfig.from_pretrained(model_id_or_path, **kwargs) | ||
| _rolling_batch_cls = get_rolling_batch_class_from_str(rolling_batch_type, model_config) |
There was a problem hiding this comment.
_rolling_batch_cls = get_rolling_batch_class_from_str(self.enable_rolling_batch, model_config)
lanking520
approved these changes
Jun 28, 2023
KexinFeng
pushed a commit
to KexinFeng/djl-serving-forked
that referenced
this pull request
Aug 16, 2023
* [python] Add rolling batch for flash attention models * Flash gptneox support * fix py tests * Set sharded to false for tp 1 * Review changes --------- Co-authored-by: sindhuso <somasundaram.sindhu@gmail.com>
KexinFeng
pushed a commit
to KexinFeng/djl-serving-forked
that referenced
this pull request
Aug 16, 2023
* [python] Add rolling batch for flash attention models * Flash gptneox support * fix py tests * Set sharded to false for tp 1 * Review changes --------- Co-authored-by: sindhuso <somasundaram.sindhu@gmail.com>
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.
Description
Brief description of what this PR is about