Bug-fixing and interface update #2
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
* wip * wip * fixed * fix missing finish reason for EOS and do length check early * optimize stop sequence handling * fixed * use list for stopped seqs * Return FinishReason.Stop with last non-empty delta * add a lock around stopped_sequences processing * remove sequence_map entries after a request is finished * fix handler
* wip * bring back assert * fix overcounted partially shared token counts
* Add sampling penalties and logit bias This PR adds the sampling penlaties of frequency penalty and presence penalty. Also it adds the logit bias. * fix * fix lint * vectorized sampling * apply code review suggestions * fix
This PR fixes the `SamplingParams.logit_bias_index` and `SamplingParams.logit_bias_value` if `SamplingParams.logit_bias` is `None`.
* remove mlc_llm dependency * use staging engine by default * wip * done
This PR fixes the frequency/presence penalty for first token.
add repetition penalty
* Update exception chaining for mlc-serve * Add more support for using contextvars
surpress noisy logging
* Fix async_connector.py * Fix * --amend
Init with tests. Server working. Major fix, serve working great. Minor fix and tests. Remove extra line. fix log_softmax use constant for number of top logprobs small clean upstream to new OpenAI API Co-authored-by: Valery Chernov <[email protected]>
vvchernov
force-pushed
the
vc/update_logprob
branch
from
97d0a1b
to
214e610
Compare
vvchernov
force-pushed
the
vc/update_logprob
branch
7 times, most recently
from
ea7a66c
to
901b16e
Compare
vvchernov
force-pushed
the
vc/update_logprob
branch
from
901b16e
to
cac70ac
Compare
zxybazh
approved these changes
Dec 31, 2023
zxybazh
merged commit 5bb9fd2
into
zxybazh:feature/2023-11-22/enable-mlc-server-logprobs
1 check passed
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.
Refactor for octoml#82 PR