-
bf08e00643
llama : refactor k-shift implementation + KV defragmentation (#5691)
Georgi Gerganov
2024-02-25 22:12:24 +02:00
-
f7625019c5
server : fix crash when system prompt is bigger than batch size (#5714)
compilade
2024-02-25 13:43:50 -05:00
-
abbabc5e51
ggml-quants : provide ggml_vqtbl1q_u8 for 64bit compatibility (#5711)
Radosław Gryta
2024-02-25 19:43:00 +01:00
-
f1a98c5254
make : fix nvcc version is empty (#5713)
kwin1412
2024-02-26 00:46:49 +08:00
-
7d548a1827
readme : add Msty to UI list (#5618)
Ashok Gelal
2024-02-25 10:57:34 -05:00
-
930b178026
server: logs - unified format and --log-format option (#5700)
Pierrick Hymbert
2024-02-25 13:50:32 +01:00
-
d52d7819b8
server: concurrency fix + monitoring - add /metrics prometheus compatible endpoint (#5708)
Pierrick Hymbert
2024-02-25 13:49:43 +01:00
-
1289408817
cmake : fix compilation for Android armeabi-v7a (#5702)
Radosław Gryta
2024-02-25 11:53:11 +01:00
-
ab336a9d5e
code : normalize enum names (#5697)
Georgi Gerganov
2024-02-25 12:09:09 +02:00
-
69917dfa55
py : fix StableLM conversion after config.json changes (#5703)
Anas Ahouzi
2024-02-25 10:54:04 +01:00
-
9e359a4f47
server: continue to update other slots on embedding concurrent request (#5699)
Pierrick Hymbert
2024-02-24 19:16:04 +01:00
-
4c4cb30736
IQ3_S: a much better alternative to Q3_K (#5676)
Kawrakow
2024-02-24 16:23:52 +02:00
-
525213d2f5
server: init functional tests (#5566)
Pierrick Hymbert
2024-02-24 12:28:55 +01:00
-
fd43d66f46
server : add KV cache quantization options (#5684)
AlpinDale
2024-02-23 19:31:54 +00:00
-
54fbcd2ce6
convert : fix missing ftype for gemma (#5690)
Jared Van Bortel
2024-02-23 13:39:14 -05:00
-
15499eb942
mpt : do not duplicate token_embd.weight on disk (#5670)
Jared Van Bortel
2024-02-22 17:05:23 -05:00
-
96633eeca1
gemma : use more bits for the token_embd.weight tensor (#5650)
Georgi Gerganov
2024-02-22 23:23:46 +02:00
-
847eedbdb2
py : add Gemma conversion from HF models (#5647)
Georgi Gerganov
2024-02-22 23:22:48 +02:00
-
7e4f339c40
ggml : always define ggml_fp16_t as uint16_t (#5666)
Georgi Gerganov
2024-02-22 23:21:39 +02:00
-
334f76fa38
sync : ggml
Georgi Gerganov
2024-02-22 23:21:05 +02:00
-
efd56b1c21
ggml : 32-bit arm compat (whisper/1891)
Georgi Gerganov
2024-02-22 18:31:40 +02:00
-
201294ae17
nix: init singularity and docker images (#5056)
Someone
2024-02-22 19:44:10 +00:00
-
5a9e2f60ba
py : minor fixes (#5668)
Georgi Gerganov
2024-02-22 20:13:25 +02:00
-
373ee3fbba
Add Gemma chat template (#5665)
Xuan Son Nguyen
2024-02-22 19:10:21 +01:00
-
4cb4d8b22d
workflows: nix: hardcode cachix ids, build unconditionally (#5663)
Someone
2024-02-22 16:32:09 +00:00
-
3a03541ced
minor : fix trailing whitespace (#5638)
Georgi Gerganov
2024-02-22 13:54:03 +02:00
-
56d03d92be
readme : update hot topics
Georgi Gerganov
2024-02-22 10:35:54 +02:00
-
a46f50747b
server : fallback to chatml, add AlphaMonarch chat template (#5628)
Xuan Son Nguyen
2024-02-22 09:33:24 +01:00
-
c5688c6250
server : clarify some params in the docs (#5640)
Alexey Parfenov
2024-02-22 08:27:32 +00:00
-
4ef245a92a
mpt : add optional bias tensors (#5638)
Dat Quoc Nguyen
2024-02-22 18:15:13 +10:00
-
973053d8b0
llama : fix loading models with shared tok_embd and output (#5651)
slaren
2024-02-22 00:42:09 +01:00
-
7c8bcc11dc
Add docs for llama_chat_apply_template (#5645)
Xuan Son Nguyen
2024-02-22 00:31:00 +01:00
-
7fe4678b02
llama : fix session save/load with quantized KV (#5649)
slaren
2024-02-21 22:52:39 +01:00
-
ba2135ccae
gemma : allow offloading the output tensor (#5646)
slaren
2024-02-21 22:18:23 +01:00
-
89febfed93
examples : do not assume BOS when shifting context (#5622)
Jared Van Bortel
2024-02-21 10:33:54 -05:00
-
5022cf242d
sync : ggml
Georgi Gerganov
2024-02-21 16:52:39 +02:00
-
1ecea255eb
server: health: fix race condition on slots data using tasks queue (#5634)
Pierrick Hymbert
2024-02-21 15:47:48 +01:00
-
a00a35cef9
readme : add LocalAI to the availables UI (#5629)
Ettore Di Giacinto
2024-02-21 15:39:10 +01:00
-
eccd7a26dd
sync : ggml (#5633)
Georgi Gerganov
2024-02-21 16:17:10 +02:00
-
c14f72db9c
readme : update hot topics
Georgi Gerganov
2024-02-21 15:39:54 +02:00
-
cc6cac08e3
llava : add --skip-unknown to 1.6 convert.py (#5632)
Daniel Bevenius
2024-02-21 14:36:57 +01:00
-
580111d42b
llama : add
gemma model (#5631)
postmasters
2024-02-21 05:08:22 -08:00
-
88c46cbdac
[SYCL] conext add name (#5624)
Meng, Hengyu
2024-02-21 17:52:06 +08:00
-
a14679cc30
IQ4_NL: 4-bit non-linear quants with blocks of 32 (#5590)
Kawrakow
2024-02-21 11:39:52 +02:00
-
6560bed3f0
server : support llava 1.6 (#5553)
CJ Pais
2024-02-20 11:07:22 -08:00
-
06bf2cf8c4
make : fix debug build with CUDA (#5616)
slaren
2024-02-20 20:06:17 +01:00
-
4ed8e4fbef
llava : add explicit instructions for llava-1.6 (#5611)
Daniel Bevenius
2024-02-20 18:30:27 +01:00
-
9c405c9f9a
Server: use llama_chat_apply_template (#5593)
Xuan Son Nguyen
2024-02-20 15:58:27 +01:00
-
5207b3fbc5
readme : update UI list (#5605)
Dane Madsen
2024-02-20 21:00:23 +11:00
-
8dbbd75754
metal : add build system support for embedded metal library (#5604)
Haoxiang Fei
2024-02-19 22:58:36 -11:00
-
c0a8c6db37
server : health endpoint configurable failure on no slot (#5594)
Pierrick Hymbert
2024-02-20 08:48:19 +01:00
-
b9111bd209
Update ggml_sycl_op_mul_mat_vec_q (#5502)
AidanBeltonS
2024-02-20 07:01:25 +00:00
-
633782b8d9
nix: now that we can do so, allow MacOS to build Vulkan binaries
Mathijs de Bruin
2024-02-13 20:28:02 +00:00
-
22f83f0c38
Enable Vulkan MacOS CI
0cc4m
2024-02-10 22:18:33 +01:00
-
bb9dcd560a
Refactor validation and enumeration platform checks into functions to clean up ggml_vk_instance_init()
0cc4m
2024-02-14 20:57:17 +01:00
-
f50db6ae0b
Add check for VK_KHR_portability_enumeration for MoltenVK support
0cc4m
2024-02-10 22:14:52 +01:00
-
d8c054517d
Add preprocessor checks for Apple devices.
Mathijs de Bruin
2024-02-06 14:39:22 +00:00
-
42f664a382
Resolve ErrorIncompatibleDriver with Vulkan on MacOS.
Mathijs de Bruin
2024-02-03 18:00:11 +00:00
-
5dde540897
Allow for Vulkan build with Accelerate.
Mathijs de Bruin
2024-02-03 17:56:46 +00:00
-
40c3a6c1e1
cuda : ignore peer access already enabled errors (#5597)
slaren
2024-02-19 23:40:26 +01:00
-
f24ed14ee0
make : pass CPPFLAGS directly to nvcc, not via -Xcompiler (#5598)
Jared Van Bortel
2024-02-19 15:54:12 -05:00
-
9d679f0fcc
examples : support minItems/maxItems in JSON grammar converter (#5039)
nopperl
2024-02-19 14:14:07 +00:00
-
1387cf60f7
llava : remove extra cont (#5587)
Georgi Gerganov
2024-02-19 15:23:17 +02:00
-
6fd413791a
llava : replace ggml_cpy with ggml_cont
slaren
2024-02-19 14:02:36 +01:00
-
337c9cbd52
sync : ggml
Georgi Gerganov
2024-02-19 14:54:21 +02:00
-
a3145bdc30
ggml-alloc : apply ggml/731
Georgi Gerganov
2024-02-19 14:53:48 +02:00
-
890559ab28
metal : option to embed MSL source into compiled binary (whisper/1842)
Didzis Gosko
2024-02-11 16:41:41 +02:00
-
d0e3ce51f4
ci : enable -Werror for CUDA builds (#5579)
Georgi Gerganov
2024-02-19 14:45:41 +02:00
-
68a6b98b3c
make : fix CUDA build (#5580)
Georgi Gerganov
2024-02-19 13:41:51 +02:00
-
70d45af0ef
readme : fix typo in README-sycl.md (#5353)
valiray
2024-02-19 02:37:10 -08:00
-
13e2c771aa
cmake : remove obsolete sycl compile flags (#5581)
Abhilash Majumder
2024-02-19 14:45:18 +05:30
-
f53119cec4
minor : fix trailing whitespace (#5538)
Georgi Gerganov
2024-02-19 10:34:10 +02:00
-
7084755396
llava : avoid changing the original BakLLaVA model (#5577)
Daniel Bevenius
2024-02-19 09:31:59 +01:00
-
4480542b22
baby-llama : allocate graphs in ggml_context (#5573)
NawafAlansari
2024-02-19 03:25:38 -05:00
-
11b12de39b
llama : add llama_chat_apply_template() (#5538)
Xuan Son Nguyen
2024-02-19 09:23:37 +01:00
-
3a9cb4ca64
cuda, metal : fix nans in soft_max (#5574)
slaren
2024-02-19 09:04:45 +01:00
-
769a716e30
readme : update (#5572)
Mirko185
2024-02-19 08:39:31 +01:00
-
f0d1fafc02
ggml : android and old glibc NUMA incompatibility bugfixes (#5557)
bmwl
2024-02-18 23:38:32 -08:00
-
a0c2dad9d4
build : pass all warning flags to nvcc via -Xcompiler (#5570)
Jared Van Bortel
2024-02-18 16:21:52 -05:00
-
14278f55d2
ggml : restore vec dot stride arg names (#5453)
Georgi Gerganov
2024-02-18 22:58:57 +02:00
-
b1de96824b
ci : fix wikitext url + compile warnings (#5569)
Georgi Gerganov
2024-02-18 22:39:30 +02:00
-
7ad554f90e
metal : fix unused warnings (#0)
Georgi Gerganov
2024-02-18 21:39:58 +02:00
-
5ee99c32f5
common, server : surface min_keep as its own parameter (#5567)
Robey Holderith
2024-02-18 11:11:16 -08:00
-
c145f8a132
server : slots monitoring endpoint (#5550)
Pierrick Hymbert
2024-02-18 18:39:57 +01:00
-
689a091bbe
sampling : do not set min_keep to n_probs (#5564)
Georgi Gerganov
2024-02-18 19:38:06 +02:00
-
f3f28c5395
cmake : fix GGML_USE_SYCL typo (#5555)
Georgi Gerganov
2024-02-18 19:17:00 +02:00
-
e75c6279d1
server : enhanced health endpoint (#5548)
Pierrick Hymbert
2024-02-18 17:31:28 +01:00
-
36376abe05
server : --n-predict option document and cap to max value (#5549)
Pierrick Hymbert
2024-02-18 17:30:09 +01:00
-
66c1968f7a
server : graceful server shutdown (#5244)
Daniel Hiltgen
2024-02-18 08:23:16 -08:00
-
1dcc3fde00
common : fix ub (#5530)
Georgi Gerganov
2024-02-18 18:21:52 +02:00
-
5d3de51f97
ggml, common, examples, tests : fixed type arguments in printf (#5528)
Herman Semenov
2024-02-18 16:20:12 +00:00
-
fc0c8d286a
llava : update surgery script to not remove tensors (#5536)
Daniel Bevenius
2024-02-18 17:19:23 +01:00
-
bd2d4e393b
1.5 bit quantization (#5453)
Kawrakow
2024-02-18 18:16:55 +02:00
-
c8e0d7efeb
flake.lock: Update
github-actions[bot]
2024-02-18 00:17:07 +00:00
-
8f1be0d42f
ggml : add ALiBi support for ggml_soft_max_ext (#5488)
Georgi Gerganov
2024-02-17 23:04:16 +02:00
-
6e4e973b26
ci : add an option to fail on compile warning (#3952)
Ananta Bastola
2024-02-17 16:03:14 -05:00
-
d250c9d61d
gitignore : update for CLion IDE (#5544)
clibdev
2024-02-17 18:28:37 +02:00
-
5bf2b94dd4
cmake : fix VULKAN and ROCm builds (#5525)
Georgi Gerganov
2024-02-16 19:05:56 +02:00
-
d2819d5577
scripts : add helpers script for bench comparing commits (#5521)
Georgi Gerganov
2024-02-16 15:14:40 +02:00
-
4cb0727698
llava : removed excess free(NULL) operation (#5531)
Herman Semenov
2024-02-16 12:43:23 +00:00