batch: migrate speculative, mtmd and server to batch_ext #29385
ngxson
merged
14 commits into
ggml-org:master
from
ngxson:xsn/llama_batch_ext_2
adapt common
d954d62f
add common_batch
38b68c5a
wip
9ce186de
wip: spec
9c8978b1
cont
70e4407d
common_speculative_process
abb6af22
server_batch to use common_batch
7c0fca3e
rm some stale calls
8ce1b957
migrate mtmd
c10598d1
handle imrope, handle return val of add()/add_embd()
5f1235fa
Merge branch 'master' into xsn/llama_batch_ext_2
924144ff
add spec zeros vector
28ce6eda
ggerganov
approved these changes
on 2026-09-27
add warning on zero fill path
2f651e66
Merge branch 'master' into xsn/llama_batch_ext_2
7bbb1843
Login to write a write a comment.
Login via GitHub