server: --n-predict option document and ensure the completion request does not exceed it #5549
Conversation
This file contains hidden or 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
b45e111 to
d112457
Compare
phymbert
commented
Feb 17, 2024
| @@ -545,6 +547,15 @@ struct llama_server_context | |||
| slot->sparams.grammar = json_value(data, "grammar", default_sparams.grammar); | |||
| slot->sparams.n_probs = json_value(data, "n_probs", default_sparams.n_probs); | |||
|
|
|||
| if (slot->n_predict > 0 && slot->params.n_predict > slot->n_predict) { | |||
There was a problem hiding this comment.
@ggerganov Should this check be moved in llama_client_slot::has_budget
4 tasks
ggerganov
approved these changes
Feb 18, 2024
jordankanter
pushed a commit
to jordankanter/llama.cpp
that referenced
this pull request
Mar 13, 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Outdated suggestions cannot be applied.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.
Context
server
--n-predictoption is supported but not documented. When endpointscompletionor the oai compatible one are called withn_predictormax_tokens, global configuration is not checked and completion tokens can exceed--n-predictserver option.It may help people to ensure request will never infinite loop for example in #3969.
Proposed changes
--n-predictoption inREADME.mdand in the server print usage.n_predictparam by the server paramsn_predictor the user input data.Open question
400when--n-predict> 0 &&data['n_predict']>--n-predictor it can be done later on