aiqubits f3586e66bc server: cap shared decision prefixes at the first output (#30710) ...

Limit the shared prefix to the first required output position of the parent and each child task.
This keeps decision output embeddings available in the final batch when questions share option tokens.

Fixes #30482

Signed-off-by: aiqubits <aiqubits@hotmail.com>

2026-10-31 19:09:48 +01:00

..

batched-bench

batch: migrate the rest of examples to llama_batch_ext (#29601)

2026-09-30 18:08:43 +02:00

cli

server: add --connect, allow p2p access to server outside LAN (#28277)

2026-10-20 00:23:14 +02:00

completion

server: add --connect, allow p2p access to server outside LAN (#28277)

2026-10-20 00:23:14 +02:00

connect

server: add --connect, allow p2p access to server outside LAN (#28277)

2026-10-20 00:23:14 +02:00

cvector-generator

batch: migrate the rest of examples to llama_batch_ext (#29601)

2026-09-30 18:08:43 +02:00

export-lora

tools : reject truncated GGUF tensor data in split and export tools (#30497)

2026-10-19 09:27:46 +03:00

fit-params

server: use port 9931 by default (#30159)

2026-10-09 10:40:03 +02:00

gguf-split

tools : reject truncated GGUF tensor data in merge (#31044)

2026-10-25 11:25:10 +02:00

imatrix

imatrix : include clocale for std::setlocale (#30079)

2026-10-07 10:58:22 +02:00

llama-bench

llama-bench: respect -fitc if bigger than required benchmark size (#28331)

2026-10-09 19:14:30 +02:00

mtmd

mtmd : resize Gemma 4 images without letterbox padding (#30922)

2026-10-31 18:55:33 +01:00

perplexity

batch: migrate the rest of examples to llama_batch_ext (#29601)

2026-09-30 18:08:43 +02:00

quantize

quantize: cap working memory size to avoid loading big tensors onto RAM (#27795)

2026-08-27 18:31:13 +02:00

results

batch: migrate the rest of examples to llama_batch_ext (#29601)

2026-09-30 18:08:43 +02:00

rpc

rpc : delete fs_path_to_utf8() (#31424)

2026-10-30 20:36:30 +01:00

server

server: cap shared decision prefixes at the first output (#30710)

2026-10-31 19:09:48 +01:00

tokenize

tokenize : drop --stdin mutual-exclusion check (#25672)

2026-07-15 18:41:51 +02:00

tts

model : add LiquidAI/LFM3-Audio-2B and LiquidAI/LFM3-Omni-2B (#31257)

2026-10-27 18:47:14 +01:00

tuning

metal : key the fa-vec tuned table by family instead of SKU (#29075)

2026-09-23 19:23:15 +08:00

ui

ui: keep chats at the bottom until the reader scrolls up (#31311)

2026-10-29 10:18:13 +02:00

CMakeLists.txt

server: add --connect, allow p2p access to server outside LAN (#28277)

2026-10-20 00:23:14 +02:00