shahondin1624

shahondin1624 pushed to main at shahondin1624/pi-extensions

2026-05-29 08:25:46 +02:00

eb1063da28 add cross-agent memory extension

3876968bfa make llama.cpp base URL configurable via settings + document live-symlink dev setup

Compare 2 commits »

shahondin1624 pushed to main at shahondin1624/pi-extensions

2026-05-29 07:34:22 +02:00

55e71b5b30 make llama.cpp base URL configurable via settings + document live-symlink dev setup

shahondin1624 pushed to main at shahondin1624/pi-extensions

2026-05-28 21:15:51 +02:00

3ddaf95610 add cross-agent memory extension

shahondin1624 pushed to main at shahondin1624/pi-extensions

2026-05-28 20:34:35 +02:00

c464f6b903 fix generation token-rate disappearing on empty completions

853cef84af fix session-handoff truncation persisting only one turn

Compare 2 commits »

shahondin1624 pushed to main at shahondin1624/pi-extensions

2026-05-27 10:42:42 +02:00

f7af660727 migrate ai-server extension from llama.cpp router to llama-swap

shahondin1624 pushed to main at shahondin1624/pi-extensions

2026-05-27 10:11:30 +02:00

40d8b30340 fix tests: replace extractCtxSize with parseCtxMapFromYaml + extractCtxFromRunningCmd

shahondin1624 pushed to main at shahondin1624/pi-extensions

2026-05-27 10:05:41 +02:00

fe82d33d94 wire real ctx-size parsing into discoverModels, fix dangling extractCtxSize import

shahondin1624 pushed to main at shahondin1624/pi-extensions

2026-05-27 09:48:23 +02:00

ede2645189 migrate ai-server extension from llama.cpp router to llama-swap

shahondin1624 pushed to master at shahondin1624/llama.cpp

2026-05-27 08:13:40 +02:00

4c66df50ca hip: fix HIP graph capture crash for FA quantized KV f16 dequant

shahondin1624 pushed to main at shahondin1624/pi-extensions

2026-05-26 14:57:10 +02:00

6a70995a98 update install script

shahondin1624 pushed to main at shahondin1624/pi-extensions

2026-05-26 12:32:36 +02:00

ff060c3e10 add session handoff implementation and tests

shahondin1624 pushed to master at shahondin1624/llama.cpp

2026-05-19 23:42:56 +02:00

a581eead32 hip: skip unsupported RDNA WMMA flash-attention cases

shahondin1624 pushed to backup-pre-upstream-rebase at shahondin1624/llama.cpp

2026-05-19 15:26:39 +02:00

shahondin1624 created branch backup-pre-upstream-rebase in shahondin1624/llama.cpp

2026-05-19 15:26:39 +02:00

shahondin1624 pushed to master at shahondin1624/llama.cpp

2026-05-19 15:26:26 +02:00

2907ee9830 turboquant: post-merge integration fixes from test validation

ddebb5ddf6 turboquant: squash-merge TheTom/llama-cpp-turboquant feature/turboquant-kv-cache

d14ce3dab4 llama : MTP clean-up (#23269)

6db130445d ui: Bump packages + address build warnings (#23300)

4b262ab662 ci : install libssl-dev (#23325)

Compare 95 commits »

shahondin1624 pushed to main at shahondin1624/pi-extensions

2026-05-19 09:18:26 +02:00

a0f95d3901 fix llama.cpp tool conversion for direct provider

shahondin1624 pushed to main at shahondin1624/pi-extensions

2026-05-19 08:35:36 +02:00

65b8231aab Update endpoint for llama.cpp

shahondin1624 pushed to main at shahondin1624/pi-extensions

2026-05-18 22:27:21 +02:00

9717a39735 remove obsolete local-llama entrypoint

4dabc93ac1 reshape pi-extensions layout to match installed extensions

Compare 2 commits »

shahondin1624 pushed to main at shahondin1624/pi-extensions

2026-05-17 23:02:47 +02:00

c44c12dd5d update readme

shahondin1624 pushed to main at shahondin1624/pi-extensions

2026-05-17 22:56:48 +02:00

01564df5be Refactor extension structure