Georgi Gerganov
1da7b76569
server : fix speculative decoding with context shift ( #10641 )
...
* server : fix speculative decoding with context shift
ggml-ci
* server : take into account speculative limits
ggml-ci
* server : add tests
2024-12-04 22:38:20 +02:00
Diego Devesa
59f4db1088
ggml : add predefined list of CPU backend variants to build ( #10626 )
...
* ggml : add predefined list of CPU backend variants to build
* update CPU dockerfiles
2024-12-04 14:45:40 +01:00
Diego Devesa
2803540814
ggml-cpu : fix HWCAP2_I8MM value ( #10646 )
2024-12-04 14:40:44 +01:00
ltoniazzi
253b7fde91
Fix HF repo commit to clone lora test models ( #10649 )
2024-12-04 10:45:48 +01:00
JFLFY2255
8d0cfd554a
llama: Support MiniCPM-1B (with & w/o longrope) ( #10559 )
2024-12-04 11:42:50 +02:00
Jeff Bolz
2759916d86
vulkan: Implement "fast divide" (mul+shift) for unary ops like copy ( #10642 )
2024-12-04 08:28:59 +01:00
Nicolò Scipione
40c6d79fb5
SYCL : Move to compile time oneMKL interface backend selection for NVIDIA backend ( #10584 )
...
* [SYCL] Move to Compile Time backend selection on oneMKL Interface for NVIDIA backend
Move to compile time selection to backend to avoid latency at run time.
Add it to all mkl gemm calls and only for NVIDIA backend.
Signed-off-by: nscipione <nicolo.scipione@codeplay.com>
* Formatting
* Address PR comments to increase readibility
---------
Signed-off-by: nscipione <nicolo.scipione@codeplay.com>
2024-12-04 09:29:20 +08:00
Wang Ran (汪然)
98036d5670
fix typo of README.md ( #10605 )
2024-12-04 02:22:50 +01:00
Frankie Robertson
cd2f37b304
Avoid using __fp16 on ARM with old nvcc ( #10616 )
2024-12-04 01:41:37 +01:00
Benson Wong
da6aac91f1
Add docs for creating a static build ( #10268 ) ( #10630 )
...
* Add notes for a static build
* Update docs/build.md
---------
Co-authored-by: Diego Devesa <slarengh@gmail.com>
2024-12-04 01:40:36 +01:00
piDack
01e6d9bb71
clip : add sycl support ( #10574 )
...
Co-authored-by: piDack <pcdack@hotmail.co>
2024-12-04 01:26:37 +01:00
Jeff Bolz
cc98896db8
vulkan: optimize and reenable split_k ( #10637 )
...
Use vector loads when possible in mul_mat_split_k_reduce. Use split_k
when there aren't enough workgroups to fill the shaders.
2024-12-03 20:29:54 +01:00
Xuan Son Nguyen
91c36c269b
server : (web ui) Various improvements, now use vite as bundler ( #10599 )
...
* hide buttons in dropdown menu
* use npm as deps manager and vite as bundler
* fix build
* fix build (2)
* fix responsive on mobile
* fix more problems on mobile
* sync build
* (test) add CI step for verifying build
* fix ci
* force rebuild .hpp files
* cmake: clean up generated files pre build
2024-12-03 19:38:44 +01:00
Georgi Gerganov
1cd3df46bd
scripts : remove amx sync
...
ggml-ci
2024-12-03 20:04:49 +02:00
Georgi Gerganov
c505471857
sync : ggml
2024-12-03 20:04:49 +02:00
mahorozte
e9e661bd59
CUDA: remove unnecessary warp reduce in FA (ggml/1032)
...
* kqmax_new_j in every thread within warp is same after operate at line 199,this reduce can be omit
* same problem in vec32
---------
Co-authored-by: ZhaoXiaoYu <zhao.xiaoyu@zte.com.cn>
2024-12-03 20:04:49 +02:00
PAB
efb6ae9630
feat: add GGML_UNARY_OP_ARGMAX Metal kernel (ggml/1019)
...
* implemented argmax kernel
* tpig -> tgpig
* change to strides
* contiguous assertions
* kernel working and tested
* argmax simd parallel implementation
* added 2 new tests for argmax in test-backend-ops
* cosmit
* added 3 tests cases for perf eval
* add test_argmax in make_test_cases_perf
* Update test-backend-ops.cpp
Co-authored-by: Diego Devesa <slarengh@gmail.com>
---------
Co-authored-by: Diego Devesa <slarengh@gmail.com>
2024-12-03 20:04:49 +02:00
PAB
667d70d170
metal : add GGML_OP_CONV_TRANSPOSE_1D kernels (ggml/1026)
...
* wip
* wip implementation f32
* kernel conv transpose 1d f32 working
* initial commit
2024-12-03 20:04:49 +02:00
Xuan Son Nguyen
3b4f2e33e2
llama : add missing LLAMA_API for llama_chat_builtin_templates ( #10636 )
2024-12-03 12:54:30 +01:00
Nikolaos Pothitos
82bca2257b
readme : add option, update default value, fix formatting ( #10271 )
...
* readme : document --no-display-prompt
* readme : update default prompt context size
* readme : remove unnecessary indentation
Indenting a line with four spaces makes Markdown treat that section as
plain text.
* readme : indent commands under bullets
* readme : indent commands in lettered list
2024-12-03 12:50:08 +02:00
Georgi Gerganov
0115df2f65
metal : small-batch mat-mul kernels ( #10581 )
...
* metal : small-batch mat-mul kernels
ggml-ci
* metal : add rest of types
ggml-ci
* metal : final adjustments
ggml-ci
* metal : add comments
ggml-ci
2024-12-03 11:52:33 +02:00
Georgi Gerganov
515d4e5372
github : minify link [no ci] (revert)
...
this doesn't work as expected
2024-12-03 11:21:43 +02:00
Georgi Gerganov
844e2e1fee
github : minify link [no ci]
2024-12-03 11:20:35 +02:00
Georgi Gerganov
70b98fadbc
server : fix default draft model parameters ( #10586 )
...
* server : force F16 KV cache for the draft model
ggml-ci
* server : fix draft params
ggml-ci
* server : various params fixes
ggml-ci
2024-12-03 11:20:00 +02:00
Xuan Son Nguyen
642330ac7c
llama : add enum for built-in chat templates ( #10623 )
...
* llama : add enum for supported chat templates
* use "built-in" instead of "supported"
* arg: print list of built-in templates
* fix test
* update server README
2024-12-02 22:10:19 +01:00
Georgi Gerganov
8648c52101
make : deprecate ( #10514 )
...
* make : deprecate
ggml-ci
* ci : disable Makefile builds
ggml-ci
* docs : remove make references [no ci]
* ci : disable swift build
ggml-ci
* docs : remove obsolete make references, scripts, examples
ggml-ci
* basic fix for compare-commits.sh
* update build.md
* more build.md updates
* more build.md updates
* more build.md updates
* Update Makefile
Co-authored-by: Diego Devesa <slarengh@gmail.com>
---------
Co-authored-by: slaren <slarengh@gmail.com>
2024-12-02 21:22:53 +02:00
haopeng
64ed2091b2
server: Add "tokens per second" information in the backend ( #10548 )
...
* add cmake rvv support
* add timings
* remove space
* update readme
* fix
* fix code
* remove empty line
* add test
---------
Co-authored-by: Xuan Son Nguyen <son@huggingface.co>
2024-12-02 14:45:54 +01:00
Akarshan Biswas
991f8aabee
SYCL: Fix and switch to GGML_LOG system instead of fprintf ( #10579 )
...
* Switched to GGML_LOG
* Fix missing semicolon
2024-12-02 15:04:11 +08:00
Georgi Gerganov
4cb003dd8d
contrib : refresh ( #10593 )
...
* contrib : refresh
* contrib : expand [no ci]
* contrib : expand test-backend-ops instructions
* contrib : add CODEOWNERS
* prs : update template to not have checkbox [no ci]
2024-12-02 08:53:27 +02:00
Juk Armstrong
917786f43d
Add mistral-v1, mistral-v3, mistral-v3-tekken and mistral-v7 chat template types ( #10572 )
...
* Templates: `mistral-v1`, `mistral-v2`, `mistral-v3`, `mistral-v3-tekken`
* Changed system message logic and added tests for all 4
* Invalid `system_message` instead of `content` fixed
* Removed tab-indented lines
* Added template code and test for `mistral-v7`
* Added all tests. Fixed bug with `tmpl == "llama2"` test.
* Replaced tabs with spaces.
* Removed `'mistral-v2'` option as no (open) models ever used it
* Removed all references to 'v2' template from comments
* Update llama.cpp
Fixed `trim_assistant_message` bug
2024-12-01 23:09:49 +01:00
Georgi Gerganov
5e1ed95583
grammars : add English-only grammar ( #10612 )
2024-12-01 21:37:54 +02:00
Wang Qin
5c7a5aa0c3
ci: add error handling for Python venv creation in run.sh ( #10608 )
2024-12-01 20:11:42 +02:00
Diego Devesa
3420909dff
ggml : automatic selection of best CPU backend ( #10606 )
...
* ggml : automatic selection of best CPU backend
* amx : minor opt
* add GGML_AVX_VNNI to enable avx-vnni, fix checks
2024-12-01 16:12:41 +01:00
alek3y
86dc11c5bc
server : bind to any port when specified ( #10590 )
2024-12-01 13:33:12 +02:00
Georgi Gerganov
6acce39710
readme : update the usage section with examples ( #10596 )
...
* readme : update the usage section with examples
* readme : more examples
2024-12-01 11:25:17 +02:00
Wang Qin
43957ef203
build: update Makefile comments for C++ version change ( #10598 )
2024-12-01 04:19:44 +01:00
Adrien Gallouët
0c39f44d70
ggml-cpu: replace AArch64 NEON assembly with intrinsics in ggml_gemv_q4_0_4x4_q8_0() ( #10567 )
...
Signed-off-by: Adrien Gallouët <angt@huggingface.co>
2024-11-30 09:13:18 -08:00
Georgi Gerganov
3e0ba0e604
readme : remove old badge
2024-11-30 10:09:21 +02:00
Georgi Gerganov
abadba05be
readme : refresh ( #10587 )
...
* readme : refresh
* readme : move section [no ci]
* readme : clarify [no ci]
* readme : fixes [no ci]
* readme : more fixes [no ci]
* readme : simplify [no ci]
* readme : clarify GGUF
2024-11-30 09:47:07 +02:00
Eve
0533e7fb38
vulkan: Dynamic subgroup size support for Q6_K mat_vec ( #10536 )
...
* subgroup 64 version with subgroup add. 15% faster
scalable version
tested for subgroup sizes 16-128
* check for subgroup multiple of 16 and greater than 16
* subgroup sizes are always a power of 2 (https://github.com/KhronosGroup/GLSL/issues/45 )
* force 16 sequential threads per block
* make 16 subgroup size a constant
2024-11-30 08:00:02 +01:00
Diego Devesa
7cc2d2c889
ggml : move AMX to the CPU backend ( #10570 )
...
* ggml : move AMX to the CPU backend
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2024-11-29 21:54:58 +01:00
Xuan Son Nguyen
b782e5c7d4
server : add more test cases ( #10569 )
...
* server : add split model test
* add test speculative
* add invalid cases
2024-11-29 21:48:56 +01:00
Robert Collins
3a8e9af402
imatrix : support combine-only ( #10492 )
...
* imatrix-combine-only idea
* ensured that behavior consistent with log
2024-11-29 19:21:37 +02:00
Diego Devesa
a3a3048e7a
cleanup UI link list ( #10577 )
...
* cleanup UI link list
* sort list alphabetically
* add missing licenses
2024-11-29 17:45:08 +01:00
Georgi Gerganov
f0678c5ff4
ggml : fix I8MM Q4_1 scaling factor conversion ( #10562 )
...
ggml-ci
2024-11-29 16:25:39 +02:00
Shupei Fan
4b3242bbea
ggml-cpu: fix typo in gemv/gemm iq4_nl_4_4 ( #10580 )
2024-11-29 14:49:02 +01:00
Alberto Cabrera Pérez
0f77aae560
sycl : offload of get_rows set to 0 ( #10432 )
2024-11-29 20:38:45 +08:00
Alberto Cabrera Pérez
266b8519ee
sycl : Reroute permuted mul_mats through oneMKL ( #10408 )
...
This PR fixes the failing MUL_MAT tests for the sycl backend.
2024-11-29 09:49:43 +00:00
Chenguang Li
938f608742
CANN: RoPE operator optimization ( #10563 )
...
* [cann] RoPE operator optimization
* [CANN]Code Formatting
---------
Co-authored-by: noemotiovon <noemotiovon@gmail.com>
2024-11-29 14:46:55 +08:00
Jeff Bolz
f095a649ec
vulkan: get the first command buffer submitted sooner ( #10499 )
...
This is an incremental improvement over #9118 to get work to the GPU a bit
sooner. The first part is to start with a smaller number of nodes before
the first submit, and ramp it up to the current 100 nodes/submit. The
second part is to reduce the dryrun overhead for all the nodes that just
need to request descriptor space.
With these changes I get around 1-2% speedup on RTX 4070 combined with my
old Haswell-era CPU.
2024-11-29 07:18:02 +01:00
