Parallel sampling eviction #157
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
masahi
force-pushed
the
parallel-sampling-eviction
branch
from
bc3dc83
to
9fb9261
Compare
Lunderberg
pushed a commit
to Lunderberg/mlc-llm
that referenced
this pull request
Jan 30, 2024
* add discord link * empty line * fix
masahi
commented
Feb 1, 2024
| f"since it has generated more than {self.max_num_batched_tokens} tokens in total" | ||
| "and currently we do not support preempting such request.", | ||
| ) | ||
| continue |
There was a problem hiding this comment.
@sunggg @elvin-n Please be aware of this limitation. Due to this, there is still a case when a parallel-sampling request is cancelled rather than preempted.
In general, we don't have a good solution for preempting a request which has generated more than max_num_batched_tokens tokens. See also #163. The easiest solution would be to stop generation at max_num_batched_tokens, but then we cannot support "unlimited" generation.
sunggg
added a commit
that referenced
this pull request
Feb 2, 2024
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Outdated suggestions cannot be applied.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.
The most interesting change is in
engine_common.pywhere I use onePrefillRequestandEvalMultiQueryRequestfor each sequence to restore cache entries for parallel-sampling requests.I couldn't find a good way to test this pragmatically. The way I tested it was to change the condition https://github.com/octoml/mlc-llm/blob/batch-serving/serve/mlc_serve/engine/engine_common.py#L327 depending on model / input to manually cause an eviction.
Ready for review @sunggg @elvin-n