PR #19833 sampling : support multiple outputs per sequence

sampling : support multiple outputs per sequence #19833

danbev wants to merge 5 commits into ggml-org:master from danbev:backend-sampling-multiple-outputs-per-seq

danbev requested a review from

ggerganov 8 days ago

danbev requested a review from

CISC 8 days ago

danbev marked this pull request as draft 8 days ago

github-actions added testing

github-actions added examples

github-actions added server

sampling : support multiple outputs per sequence

1138d5c2

llama : add n_sampling_outputs_max cparam

1e8c02aa

llama : enable static graph for multiple sampling outputs per sequence

765998f2

server : enable backend sampling for multiple outputs per sequence

2235b4be

sampling : add clamping to backend dist sampler

5c92c76e

danbev force pushed from 6c6b36bc to 5c92c76e 5 days ago

danbev marked this pull request as ready for review 5 days ago

danbev requested a review from

ngxson 5 days ago

Reviewers

ggerganov

CISC

ngxson

Assignees

No one assigned

Labels

testing examples server

Milestone

No milestone