]>
git.djapps.eu Git - pkg/ggml/sources/llama.cpp/log
cmdr2 [Thu, 10 Apr 2025 12:23:08 +0000 (17:53 +0530)]
ggml: don't include arm_neon.h when using CUDA 12 with ARM Neon (ggml/1187)
fix #1186
Diego Devesa [Wed, 9 Apr 2025 10:32:13 +0000 (12:32 +0200)]
ggml : add bilinear upscale support (ggml/1185)
Diego Devesa [Wed, 9 Apr 2025 10:31:34 +0000 (12:31 +0200)]
ggml : add more generic custom op, remove deprecated custom ops (ggml/1183)
* ggml : add more generic ggml_custom op
* ggml : remove deprecated custom ops
Georgi Gerganov [Thu, 10 Apr 2025 20:59:01 +0000 (23:59 +0300)]
scripts : fix sync-ggml-am.sh
Xuan-Son Nguyen [Thu, 10 Apr 2025 20:57:16 +0000 (22:57 +0200)]
llava : introduce libmtmd (#12849)
* wip llava2
* migrated gemma3 to llava2
* add timings
* correct pre/postfix
* fix missing include
* fix compilation unused var warn
* update llava2_tokenize
* change name llava2 --> mtmd
* improve api
* refine helpers
* Update examples/llava/mtmd.cpp
Co-authored-by: Georgi Gerganov <redacted>
---------
Co-authored-by: Georgi Gerganov <redacted>
Xuan-Son Nguyen [Thu, 10 Apr 2025 15:24:44 +0000 (17:24 +0200)]
convert : ability to lazy-load safetensors remotely without downloading to disk (#12820)
* gguf util : add SafetensorRemote
* fix style
* convert: add --remote option
* convert : allow using lazy remote tensors
It's a bit slow for now since everything is blocking and single-threaded.
* correct metadata.name
* small style fix
* support HF_TOKEN
* convert : use writeable buffer for remote lazy tensors
* convert : fix flake8 lint regarding lamdba assigment
* multithreaded download
* multithread: print debug
* fix style
* Revert "multithreaded download"
This reverts commit
42fc895ace385edc972ad819c76c704aeea61791 .
* bring back _get_request_headers
---------
Co-authored-by: Francis Couture-Harpin <redacted>
Chenguang Li [Thu, 10 Apr 2025 00:51:52 +0000 (08:51 +0800)]
CANN: Support more ops (#12841)
* [CANN]Support Opt LOG && MEAN && PAD_REFLECT_1D
* [CANN]Support COUNT_EQUAL && STEP && SGN
* [CANN]codestyle adjustment
* [CANN]codestyle adjustment
---------
Signed-off-by: noemotiovon <redacted>
Prajwal B Mehendarkar [Wed, 9 Apr 2025 23:18:01 +0000 (04:48 +0530)]
Fixes #12823 (#12830)
* Including limits file on AIX
* Fixes #12823
Rudi Servo [Wed, 9 Apr 2025 23:17:12 +0000 (23:17 +0000)]
docker : added all CPU to GPU images (#12749)
Piotr Kubaj [Wed, 9 Apr 2025 23:00:34 +0000 (23:00 +0000)]
ggml-cpu-impl.h: do not redefine bool on POWER9 (#12856)
error: unknown type name '_Bool'
Piotr Kubaj [Wed, 9 Apr 2025 23:00:25 +0000 (23:00 +0000)]
ggml-impl.h: fix build on POWER9 (#12855)
error: ISO C++17 does not allow 'register' storage class specifier
Bo Zheng [Wed, 9 Apr 2025 09:47:36 +0000 (17:47 +0800)]
llama : Support Qwen3 and Qwen3MoE (#12828)
* add qwen3 & qwen3moe support.
* fix
---------
Co-authored-by: bozheng-hit <redacted>
R0CKSTAR [Wed, 9 Apr 2025 09:22:30 +0000 (17:22 +0800)]
musa: enable freediskspace for docker image build (#12839)
Signed-off-by: Xiaodong Ye <redacted>
Romain Biessy [Wed, 9 Apr 2025 09:22:04 +0000 (11:22 +0200)]
sycl: update documentation to use -no-cnv (#12845)
Plamen Minev [Wed, 9 Apr 2025 08:11:11 +0000 (11:11 +0300)]
ci: detach common from the library (#12827)
* fix: detach common from the library
* fix: building chat test template
Xuan-Son Nguyen [Wed, 9 Apr 2025 08:09:53 +0000 (10:09 +0200)]
clip : do not print ftype (#12832)
Georgi Gerganov [Wed, 9 Apr 2025 07:54:42 +0000 (10:54 +0300)]
readme : add rpc backend (#12842)
Chenguang Li [Wed, 9 Apr 2025 06:04:14 +0000 (14:04 +0800)]
CANN: Support Opt CONV_TRANSPOSE_1D and ELU (#12786)
* [CANN] Support ELU and CONV_TRANSPOSE_1D
* [CANN]Modification review comments
* [CANN]Modification review comments
* [CANN]name adjustment
* [CANN]remove lambda used in template
* [CANN]Use std::func instead of template
* [CANN]Modify the code according to the review comments
---------
Signed-off-by: noemotiovon <redacted>
Jeff Bolz [Wed, 9 Apr 2025 05:25:08 +0000 (00:25 -0500)]
vulkan: In coopmat2 mmq, load q4_k/q5_k scales through shared memory (#12833)
q4_k and q5_k had a lot of redundant global loads where the same 16B of
scale information is repeatedly loaded and decoded during each loop iteration.
This change restructures the loops to more explicitly iterate over whole
blocks in the outer loop (with unrolled inner loop) and to copy/decode the
scale data into shared memory once at the start of each outer loop. The copy
is pipelined so the scale load from global memory is relatively cheap.
This improves q4_k/q5_k model prompt processing performance by around 5-7%.
I briefly tried applying this to q6_k and q4_0, and it didn't help for q6_k
and hurt for q4_0.
The big "else" path in mul_mm_cm2.comp that had all the clamped/unclamped
variants isn't used as often as it originally was (e.g. due to the padded_N
change), so I trimmed it down to offset some of the new complexity of the
semi-manual loop unrolling.
Jeff Bolz [Wed, 9 Apr 2025 05:12:57 +0000 (00:12 -0500)]
vulkan: Use fp16 for the flash attention P*V multiplication (#12783)
This is consistent with the ggml-cuda behavior and the mul_mat fallback.
Sigbjørn Skjæret [Tue, 8 Apr 2025 21:21:31 +0000 (23:21 +0200)]
cuda : add f32 to bf16 copy op (#12806)
This allows BF16 KV-cache on CUDA.
Matt Clayton [Tue, 8 Apr 2025 20:01:58 +0000 (16:01 -0400)]
llava: improve clip_ctx destructor to not memleak load_image_size (#12834)
Georgi Gerganov [Tue, 8 Apr 2025 16:54:51 +0000 (19:54 +0300)]
llama : fix FA when KV cache is not used (i.e. embeddings) (#12825)
* ggml : FA supports F32 V
* graph : cast KV to F16 when the KV cache is not used
ggml-ci
* server : add test that exercises embeddings with FA enabled
ggml-ci
Xuan-Son Nguyen [Tue, 8 Apr 2025 16:37:06 +0000 (18:37 +0200)]
server : fix thread.join() on exit (#12831)
dm4 [Tue, 8 Apr 2025 13:49:13 +0000 (21:49 +0800)]
llava: add more helper functions to check projector types in clip context (#12824)
Signed-off-by: dm4 <redacted>
Prajwal B Mehendarkar [Tue, 8 Apr 2025 12:30:59 +0000 (18:00 +0530)]
arg : Including limits file on AIX (#12822)
characharm [Tue, 8 Apr 2025 09:14:59 +0000 (14:14 +0500)]
server : webui : Improve Chat Input with Auto-Sizing Textarea (#12785)
* Update ChatScreen.tsx
* useAutosizeTextarea.ts
useAutosizeTextarea to encapsulate the logic.
* Implement responsive auto-sizing chat textarea
Replaces the manual textarea resizing with an automatic height adjustment based on content.
- `useChatTextarea` hook to manage textarea state and auto-sizing logic via refs, preserving the optimization
- Textarea now grows vertically up to a maximum height (`lg:max-h-48`) on large screens (lg breakpoint and up).
- Disables auto-sizing and enables manual vertical resizing (`resize-vertical`) on smaller screens for better mobile usability.
- Aligns the "Send" button to the bottom of the textarea (`items-end`) for consistent positioning during resize.
* -update compressed index.html.gz after npm run build
-refactor: replace OptimizedTextareaValue with AutosizeTextareaApi in VSCode context hook
* chore: normalize line endings to LF
refactor: AutosizeTextareaApi -> chatTextareaApi
* refactor: Rename interface to PascalCase
---------
Co-authored-by: Xuan Son Nguyen <redacted>
Neo Zhang Jianyu [Tue, 8 Apr 2025 07:03:21 +0000 (15:03 +0800)]
Revert "sycl:remove redundant memcopy in function ggml_backend_sycl_buffer_set_tensor" (#12812)
* Revert "sycl: remove redundant memcopy in function ggml_backend_sycl_buffer_s…"
This reverts commit
518a01480eb3a7c80a4951b430db9dee55428310 .
* Update ggml/src/ggml-sycl/ggml-sycl.cpp
* Update ggml/src/ggml-sycl/ggml-sycl.cpp
* rm tail space
compilade [Tue, 8 Apr 2025 07:03:07 +0000 (03:03 -0400)]
gguf-py : support lazy tensor splitting (#12809)
* gguf-py : support lazy tensor splitting
Splitting usually involves returning tuples of tensors,
which need to be handled properly to avoid early eager evaluation.
* gguf-py : fix flake8 lint
Xuan-Son Nguyen [Mon, 7 Apr 2025 21:06:44 +0000 (23:06 +0200)]
llama : Support llama 4 text-only (#12791)
* llama4 conversion
* initial support, no chat template
* clean up a bit
* fix tokenizer conversion
* correct hparams
* try this
* fix shexp
* ffn_inp_normed
* chat template
* clean up model conversion
* add_bos
* add scale_before_ffn
* fix order
* weight_before_ffn
* llm_graph_input_attn_temp
* add chunk attn mask
* build_inp_attn_scale()
* add comment about ggml_repeat
* clarify comments
* fix build
lhez [Mon, 7 Apr 2025 20:22:54 +0000 (13:22 -0700)]
opencl: better identify Adreno GPU (#12760)
stduhpf [Mon, 7 Apr 2025 15:47:08 +0000 (17:47 +0200)]
hellaswag: display estimated score confidence interval (#12797)
Georgi Gerganov [Mon, 7 Apr 2025 10:18:07 +0000 (13:18 +0300)]
cuda : fix HIP and MUSA BF16 (#0)
ggml-ci
Georgi Gerganov [Mon, 7 Apr 2025 09:32:39 +0000 (12:32 +0300)]
sync : ggml
ggml-ci
Georgi Gerganov [Mon, 7 Apr 2025 09:25:15 +0000 (12:25 +0300)]
ggml : simplify Arm fp16 CPU logic (ggml/1177)
* ggml : simlpify Arm fp16 CPU logic
ggml-ci
* cont : bring back CUDA/MUSA checks
ggml-ci
Sigbjørn Skjæret [Fri, 4 Apr 2025 19:05:12 +0000 (21:05 +0200)]
CUDA: don't convert BF16 weights to FP32 (ggml/1174)
* add bf16 support
* use convert_from_bf16_cuda instead of convert_unary_cuda for f32
* revert
7ec5085
* move functionality into convert_unary with constexpr
cmdr2 [Wed, 2 Apr 2025 12:16:16 +0000 (17:46 +0530)]
cpu: move all the operators into a separate c++ file (except mul_mat) (ggml/1167)
* cpu: refactor SIMD mappings and vectorized op functions into separate files
* Fix warning for ggml_float to float
* Fix warnings
* cpu: move all the operations (except mul_mat) to a separate c++ file
* fix whitespace
* Update ggml/src/ggml-cpu/vec.h
Co-authored-by: Diego Devesa <redacted>
* Fix PR comments - use GGML_UNUSED, use cassert in ops.cpp
* Reverse the order of import for ops.h and vec.h, to match what was present in ggml-cpu.c previously
---------
Co-authored-by: Diego Devesa <redacted>
zhouwg [Mon, 7 Apr 2025 15:22:57 +0000 (23:22 +0800)]
sycl: remove redundant memcopy in function ggml_backend_sycl_buffer_set_tensor (#12734)
Xuan-Son Nguyen [Mon, 7 Apr 2025 12:37:28 +0000 (14:37 +0200)]
ci : no curl on ggml-ci (#12796)
Xuan-Son Nguyen [Mon, 7 Apr 2025 11:35:19 +0000 (13:35 +0200)]
cmake : enable curl by default (#12761)
* cmake : enable curl by default
* no curl if no examples
* fix build
* fix build-linux-cross
* add windows-setup-curl
* fix
* shell
* fix path
* fix windows-latest-cmake*
* run: include_directories
* LLAMA_RUN_EXTRA_LIBS
* sycl: no llama_curl
* no test-arg-parser on windows
* clarification
* try riscv64 / arm64
* windows: include libcurl inside release binary
* add msg
* fix mac / ios / android build
* will this fix xcode?
* try clearing the cache
* add bunch of licenses
* revert clear cache
* fix xcode
* fix xcode (2)
* fix typo
zhouwg [Mon, 7 Apr 2025 11:34:14 +0000 (19:34 +0800)]
CANN: fix typo in ggml-cann (#12733)
hipudding [Mon, 7 Apr 2025 09:10:36 +0000 (17:10 +0800)]
CANN: Refactor to reduce duplicate code (#12731)
* CANN: Refactor to reduce duplicate code
* CANN: fix review comment
R0CKSTAR [Sun, 6 Apr 2025 13:23:54 +0000 (21:23 +0800)]
musa: fix compilation warnings in mp_22/31 (#12780)
Signed-off-by: Xiaodong Ye <redacted>
Jeff Bolz [Sun, 6 Apr 2025 09:03:47 +0000 (04:03 -0500)]
vulkan: fix NaN issue in flash attention shader (#12776)
Use -FLT_MAX/2 rather than -inf as the initial value for computing the maximum.
Jeff Bolz [Sun, 6 Apr 2025 08:47:13 +0000 (03:47 -0500)]
vulkan: Use unclamped loads for flash attention mask (#12720)
nem1 must be a multiple of GGML_KQ_MASK_PAD, and GGML_KQ_MASK_PAD is a multiple
of the number of rows in the matrix. The KV dim is a multiple of the number of
columns for the aligned shader.
0cc4m [Sat, 5 Apr 2025 16:04:03 +0000 (18:04 +0200)]
Vulkan: Tune Vulkan mmq int dot shader for performance (#12767)
Sergey Fedorov [Sat, 5 Apr 2025 15:46:00 +0000 (23:46 +0800)]
common : fix includes in arg.cpp and gemma3-cli.cpp (#12766)
* arg.cpp: add a missing include
* gemma3-cli.cpp: fix cinttypes include
Xuan-Son Nguyen [Sat, 5 Apr 2025 15:17:40 +0000 (17:17 +0200)]
clip : refactor clip_init, add tests (#12757)
* refactor clip_init
* fix loading file
* fix style
* test ok
* better test with report
* add missing headers
* clarify
* add KEY_MM_PATCH_MERGE_TYPE
* remove bool has_* pattern
* Apply suggestions from code review
Co-authored-by: Georgi Gerganov <redacted>
* Update examples/llava/clip.cpp
Co-authored-by: Georgi Gerganov <redacted>
* use ggml_soft_max_ext
* refactor logging system
* add minicpm-v-o 2.6 for testing
* use nullptr everywhere
* fix Yi-VL model
---------
Co-authored-by: Georgi Gerganov <redacted>
エシュナヴァリシア [Sat, 5 Apr 2025 13:31:42 +0000 (21:31 +0800)]
common: custom hf endpoint support (#12769)
* common: custom hf endpoint support
Add support for custom huggingface endpoints via HF_ENDPOINT environment variable
You can now specify a custom huggingface endpoint using the HF_ENDPOINT environment variable when using the --hf-repo flag, which works similarly to huggingface-cli's endpoint configuration.
Example usage:
HF_ENDPOINT=https://hf-mirror.com/ ./bin/llama-cli --hf-repo Qwen/Qwen1.5-0.5B-Chat-GGUF --hf-file qwen1_5-0_5b-chat-q2_k.gguf -p "The meaning to life and the universe is"
The trailing slash in the URL is optional:
HF_ENDPOINT=https://hf-mirror.com ./bin/llama-cli --hf-repo Qwen/Qwen1.5-0.5B-Chat-GGUF --hf-file qwen1_5-0_5b-chat-q2_k.gguf -p "The meaning to life and the universe is"
* Update common/arg.cpp
readability Improvement
Co-authored-by: Xuan-Son Nguyen <redacted>
* Apply suggestions from code review
---------
Co-authored-by: ベアトリーチェ <redacted>
Co-authored-by: Xuan-Son Nguyen <redacted>
Olivier Chafik [Fri, 4 Apr 2025 20:16:39 +0000 (13:16 -0700)]
sync: minja (#12739)
* sync: minja
https://github.com/google/minja/pull/57
* fix json include
Georgi Gerganov [Fri, 4 Apr 2025 18:48:10 +0000 (21:48 +0300)]
kv-cache : simplify + fix warning for recurrent models (#12756)
ggml-ci
bandoti [Fri, 4 Apr 2025 17:05:12 +0000 (14:05 -0300)]
ci: add Linux cross-compile build (#12428)
Nauful Shaikh [Fri, 4 Apr 2025 14:09:52 +0000 (09:09 -0500)]
server : webui : Upgrade daisyui, tailwindcss. (#12735)
* Upgrade daisyui, tailwindcss.
* Switch to all themes.
* Revert a change.
* Update formatting.
* Install packages before npm build.
* Revert "Install packages before npm build."
This reverts commit
336c5147e614e60993162794ba9d9d4629a916f8 .
* Add index.html.gz
* run build
---------
Co-authored-by: Xuan Son Nguyen <redacted>
nick huang [Fri, 4 Apr 2025 14:09:12 +0000 (22:09 +0800)]
gguf-split : --merge now respects --dry-run option (#12681)
* gguf-split now respects dry-run option
* removing trailing space
Nicolò Scipione [Fri, 4 Apr 2025 14:00:46 +0000 (16:00 +0200)]
sycl: allow ggml-sycl configuration and compilation using Visual Studio project/solution (#12625)
Ronny Brendel [Fri, 4 Apr 2025 13:12:40 +0000 (15:12 +0200)]
cmake: fix ggml-shaders-gen compiler paths containing spaces (#12747)
fixes error for compiler paths with spaces
Daniel Bevenius [Fri, 4 Apr 2025 08:24:12 +0000 (10:24 +0200)]
docs : add XCFramework section to README.md [no ci] (#12746)
This commit adds a new section to the README.md file, detailing the
usage of the XCFramework.
The motivation for this is that it might not be immediately clear to
users how to use the XCFramework in their projects and hopefully this
will help.
Jeff Bolz [Fri, 4 Apr 2025 05:54:35 +0000 (00:54 -0500)]
vulkan: Hybrid waitForFences/getFenceStatus to reduce fence latency (#12630)
There seems to be a bubble waking up from waitForFences, which costs a few
percent performance and also increased variance in performance. This change
inserts an "almost_ready" fence when the graph is about 80% complete and we
waitForFences for the almost_ready fence and then spin (with _mm_pauses) waiting
for the final fence to be signaled.
Jeff Bolz [Fri, 4 Apr 2025 05:53:20 +0000 (00:53 -0500)]
vulkan: set cmake minimum and project name in vulkan-shaders (#12744)
lhez [Fri, 4 Apr 2025 05:18:17 +0000 (22:18 -0700)]
opencl: update doc for OpenCL (#12702)
* opencl: add OpenCL to build.md
* opencl: remove fixed issue/TODO
* opencl: add link to OPENCL.md
* opencl: update doc - refine tools requirement for Windows 11 arm64
Gaurav Garg [Thu, 3 Apr 2025 16:20:29 +0000 (21:50 +0530)]
CUDA: Prefer vector flash decoding kernel for Gemma models (#12738)
* Prefer vector flash decoding kernel for Gemma models
Vector flash decoding kernel was not being picked for models with head dimension 256. Gemma models are in this category.
Removing this limit improves e2e performance by upto 12% in gen phase throughput for Gemm models.
* Update ggml/src/ggml-cuda/fattn.cu
Co-authored-by: Johannes Gäßler <redacted>
---------
Co-authored-by: Johannes Gäßler <redacted>
yumeyao [Thu, 3 Apr 2025 15:32:54 +0000 (23:32 +0800)]
vocab : use string_view::find() to avoid unnecessary looking up beyond the fragment range (#12706)
Jeff Bolz [Thu, 3 Apr 2025 15:08:26 +0000 (10:08 -0500)]
vulkan: Fix missing cmake logic for dot product extension (#12721)
Atharva Dubey [Thu, 3 Apr 2025 12:12:39 +0000 (13:12 +0100)]
ci : add env variable in ggml-ci and document the same in SYCL.md (#12736)
R0CKSTAR [Thu, 3 Apr 2025 11:51:35 +0000 (19:51 +0800)]
sync : minja (inclusionAI/Ling) and update tests (#12699)
Signed-off-by: Xiaodong Ye <redacted>
a3sh [Thu, 3 Apr 2025 07:32:55 +0000 (15:32 +0800)]
fix MUSA compiler warning (#12704)
* fix MUSA compiler warning
* replace (void) with GGML_UNUSED
Chenguang Li [Thu, 3 Apr 2025 07:18:08 +0000 (15:18 +0800)]
CANN: Support operator SIN COS ARGMAX (#12709)
* [CANN]support sin cos argmax
Signed-off-by: noemotiovon <redacted>
* [CANN]codestyle adjustment
Signed-off-by: noemotiovon <redacted>
* [CANN]Remove redundant code
Signed-off-by: noemotiovon <redacted>
---------
Signed-off-by: noemotiovon <redacted>
Co-authored-by: noemotiovon <redacted>
Alan Gray [Thu, 3 Apr 2025 01:31:15 +0000 (02:31 +0100)]
Simplify and improve CUDA graphs through use of indirect copy pointers (#9017)
* CUDA: Simplify and improve CUDA graphs through use of indirect copy pointers
Previously there was complexity in the CUDA graphs implementation due
frequently changing parameters to copy kernels associated with K and V
cache pointers. This patch simplifies by using indirection to avoid
such parameters frequently changing, avoiding the need for frequent
graph updates.
Fixes #12152
* Addressed comments
* fix HIP builds
* properly sync to stream
* removed ggml_cuda_cpy_fn_ptrs
* move stream sync before free
* guard to only use indirection with graphs
* style fixes
* check for errors
---------
Co-authored-by: slaren <redacted>
hipudding [Thu, 3 Apr 2025 00:49:51 +0000 (08:49 +0800)]
CANN: Fix failed test cases (#12708)
* CANN: Fix memory waste in aclnn_tensor
* CANN: fix backend ops fail
* CANN: fix acl_tensor memory alloc.
* CANN: format
* CANN: remove trailing whitespace
lhez [Thu, 3 Apr 2025 00:01:42 +0000 (17:01 -0700)]
opencl: use `max_alloc_size` in backend ctx instead of querying again (#12705)
Jeff Bolz [Wed, 2 Apr 2025 19:25:08 +0000 (14:25 -0500)]
vulkan: Implement split_k for coopmat2 flash attention. (#12627)
When using group query attention, we have one workgroup per KV batch and this
can be very few workgroups (e.g. just 8 in some models). Enable split_k to
spread the work across SMs. This helps a lot when the KV cache is large.
bandoti [Wed, 2 Apr 2025 17:56:26 +0000 (14:56 -0300)]
cmake: remove caching from vulkan coopmat checks (#12719)
Jeff Bolz [Wed, 2 Apr 2025 17:40:32 +0000 (12:40 -0500)]
vulkan: Implement grouped query attention in the coopmat2 FA shader (#12559)
When adjacent batches of Q share the same batches of K/V, batch them into
the same workgroup. For example, when:
dst(128,32,1,1) = FA(q(128,1,32,1), k(128,16640,8,1), v(128,16640,8,1))
previously we would run 32 workgroups computing 1 result each, now we will
run 8 workgroups computing 4 results each.
This doesn't directly translate to better performance (at least when you have
>=32 SMs), but in a subsequent change I'll enable split_k which will scale much
better with 4x fewer workgroups.
0cc4m [Wed, 2 Apr 2025 17:12:30 +0000 (19:12 +0200)]
Vulkan: Fix mmq int dot float cache size (#12722)
Georgi Gerganov [Wed, 2 Apr 2025 13:38:54 +0000 (16:38 +0300)]
model : print tensor size during load (#12711)
* model : print tensor size during load
* cont : fix units MB -> MiB
Co-authored-by: Diego Devesa <redacted>
---------
Co-authored-by: Diego Devesa <redacted>
Diego Devesa [Wed, 2 Apr 2025 12:52:01 +0000 (14:52 +0200)]
llama : add option to override model tensor buffers (#11397)
* llama : add option to override tensor buffers
* ggml : fix possible underflow in ggml_nbytes
Georgi Gerganov [Wed, 2 Apr 2025 11:32:59 +0000 (14:32 +0300)]
llama : refactor kv cache guard (#12695)
* llama : refactor kv cache guard
ggml-ci
* cont : fix comment [no ci]
* llama : fix kv_cache restore logic
ggml-ci
* context : simplify kv cache updates
ggml-ci
* cont : better name [no ci]
* llama : fix llama_decode return code when could not find KV slot
ggml-ci
* context : change log err -> warn [no ci]
* kv-cache : add comment + warning
Sigbjørn Skjæret [Wed, 2 Apr 2025 09:21:48 +0000 (11:21 +0200)]
vocab : BailingMoE : change possessive quantifiers to greedy (#12677)
Xuan-Son Nguyen [Wed, 2 Apr 2025 07:58:34 +0000 (09:58 +0200)]
common : remove json.hpp from common.cpp (#12697)
* common : remove json.hpp from common.cpp
* fix comment
Chenguang Li [Wed, 2 Apr 2025 07:22:13 +0000 (15:22 +0800)]
[CANN] get_rows and dup optimization (#12671)
* [CANN]get_rows and dup optimization.
Co-authored-by: hipudding <redacted>
Signed-off-by: noemotiovon <redacted>
* [CANN]GET_ROWS and CPY/DUP optimization
Co-authored-by: hipudding <redacted>
Signed-off-by: noemotiovon <redacted>
* [CANN]code style adjustment
Signed-off-by: noemotiovon <redacted>
* [CANN]code style adjustment
Signed-off-by: noemotiovon <redacted>
* [CANN]code style adjustment
Signed-off-by: noemotiovon <redacted>
* [CANN]code style adjustment
Signed-off-by: noemotiovon <redacted>
---------
Signed-off-by: noemotiovon <redacted>
Co-authored-by: noemotiovon <redacted>
Co-authored-by: hipudding <redacted>
Xuan-Son Nguyen [Tue, 1 Apr 2025 21:44:05 +0000 (23:44 +0200)]
common : refactor downloading system, handle mmproj with -hf option (#12694)
* (wip) refactor downloading system [no ci]
* fix all examples
* fix mmproj with -hf
* gemma3: update readme
* only handle mmproj in llava example
* fix multi-shard download
* windows: fix problem with std::min and std::max
* fix 2
Junil Kim [Tue, 1 Apr 2025 16:54:34 +0000 (01:54 +0900)]
opencl : fix memory allocation size (#12649)
issue:
https://github.com/CodeLinaro/llama.cpp/pull/17#issuecomment-
2760611283
This patch fixes the memory allocation size
not exceeding the maximum size of the OpenCL device.
jklincn [Tue, 1 Apr 2025 12:54:28 +0000 (20:54 +0800)]
llama : use LLM_KV_GENERAL_FILE_TYPE instead of gguf_find_key (#12672)
Sigbjørn Skjæret [Tue, 1 Apr 2025 12:37:13 +0000 (14:37 +0200)]
convert : BailingMoE : fix qkv split when head_dim is 0 (#12687)
NOTE: Ling-lite-base is broken, see https://huggingface.co/inclusionAI/Ling-lite-base/discussions/2
Georgi Gerganov [Tue, 1 Apr 2025 11:57:19 +0000 (14:57 +0300)]
metal : use F32 prec in FA kernels (#12688)
* metal : use F32 prec in FA kernels
ggml-ci
* cont : fix FA vec kernel
ggml-ci
R0CKSTAR [Tue, 1 Apr 2025 11:12:53 +0000 (19:12 +0800)]
Fix clang warning in gguf_check_reserved_keys (#12686)
* Fix clang warning in gguf_check_reserved_keys
Signed-off-by: Xiaodong Ye <redacted>
* Fix typo
Signed-off-by: Xiaodong Ye <redacted>
---------
Signed-off-by: Xiaodong Ye <redacted>
Wagner Bruna [Tue, 1 Apr 2025 09:38:07 +0000 (06:38 -0300)]
vulkan: fix build when glslc doesn't support coopmat (#12683)
Romain Biessy [Tue, 1 Apr 2025 08:24:29 +0000 (10:24 +0200)]
SYCL: Rename oneMKL to oneMath (#12192)
* Rename oneMKL Interface to oneMath
* Use oneMath for Intel vendor
* Rename occurences to mkl
* clang-format
* Silence verbose warnings
* Set oneMath HIP_TARGETS
* Fix silence warnings
* Remove step to build oneMath from build instructions
* Use fixed oneMath version
* Remove INTEL_CPU
* Fold CMake oneDNN conditions
* Use Intel oneMKL for Intel devices
* Improve CMake message
* Link against MKL::MKL_SYCL::BLAS only
* Move oneMath documentation to Nvidia and AMD sections
Akarshan Biswas [Tue, 1 Apr 2025 08:11:39 +0000 (13:41 +0530)]
SYCL: switch to SYCL namespace (#12674)
Sigbjørn Skjæret [Mon, 31 Mar 2025 21:09:48 +0000 (23:09 +0200)]
convert : BailingMoE : avoid setting rope_dim to 0 (#12678)
Daniel Bevenius [Mon, 31 Mar 2025 16:40:56 +0000 (18:40 +0200)]
vocab : add special infill tokens for CodeLlama (#11850)
* vocab : add special infill tokens for CodeLlama
The commit adds the following special tokens for CodeLlama infill:
- `▁<PRE>`
- `▁<SUF>`
- `▁<MID>`
The motivation for this is that currently the infill example uses
CodeLlama as a suggested model. But when using this model the following
error is generated:
```console
/llama.cpp-debug/examples/infill/infill.cpp:165: GGML_ASSERT(llama_vocab_fim_pre(vocab) >= 0) failed
Could not attach to process. If your uid matches the uid of the target
process, check the setting of /proc/sys/kernel/yama/ptrace_scope, or try
again as the root user. For more details, see /etc/sysctl.d/10-ptrace.conf
ptrace: Operation not permitted.
No stack.
The program is not being run.
305251 Aborted (core dumped)
./build/bin/llama-infill -t 10 -ngl 0 -m models/codellama-13b.Q5_K_S.gguf \
-c 4096 --temp 0.7 --repeat_penalty 1.1 -n 20 \
--in-prefix "def helloworld():\n print(\"hell" \
--in-suffix "\n print(\"goodbye world\")\n "
```
* squash! vocab : add special infill tokens for CodeLlama
Add _<EOT> as well.
a3sh [Mon, 31 Mar 2025 16:05:13 +0000 (00:05 +0800)]
ggml : faster ssm scan (#10558)
* faster ssm_scan
* delete unused commnet
* clang format
* add space
* modify unnecessary calculations
* faster ssm conv implementatioin
* modify file name with dash
Sigbjørn Skjæret [Mon, 31 Mar 2025 14:36:25 +0000 (16:36 +0200)]
convert : Qwerky : use lora_rank_tokenshift and lora_rank_decay if present (#12667)
0cc4m [Mon, 31 Mar 2025 12:37:01 +0000 (14:37 +0200)]
Vulkan: Add DP4A MMQ and Q8_1 quantization shader (#12135)
* Vulkan: Add DP4A MMQ and Q8_1 quantization shader
* Add q4_0 x q8_1 matrix matrix multiplication support
* Vulkan: Add int8 coopmat MMQ support
* Vulkan: Add q4_1, q5_0 and q5_1 quants, improve integer dot code
* Add GL_EXT_integer_dot_product check
* Remove ggml changes, fix mmq pipeline picker
* Remove ggml changes, restore Intel coopmat behaviour
* Fix glsl compile attempt when integer vec dot is not supported
* Remove redundant code, use non-saturating integer dot, enable all matmul sizes for mmq
* Remove redundant comment
* Fix integer dot check
* Fix compile issue with unsupported int dot glslc
* Update Windows build Vulkan SDK version
Georgi Gerganov [Mon, 31 Mar 2025 12:05:30 +0000 (15:05 +0300)]
cmake : fix whitespace (#0)
Georgi Gerganov [Mon, 31 Mar 2025 11:59:21 +0000 (14:59 +0300)]
sync : ggml
ggml-ci
Sandro Hanea [Mon, 31 Mar 2025 10:44:36 +0000 (12:44 +0200)]
cmake: improve Vulkan cooperative matrix support checks (whisper/2966)
Co-authored-by: Sandro Hanea <redacted>
Sigbjørn Skjæret [Mon, 31 Mar 2025 09:28:30 +0000 (11:28 +0200)]
llava : proper description fix (#12668)
Akarshan Biswas [Mon, 31 Mar 2025 09:25:24 +0000 (14:55 +0530)]
SYCL: Remove misleading ggml_sycl_op_flatten function (#12387)
* SYCL: Remove misleading ggml_sycl_op_flatten function
* remove trailing whitespace
* Fix L2 norm from rebase
* remove try catch block from element_wise.cpp
* remove comment from common.hp
* ggml-sycl.cpp: Add try catch sycl::exception block in compute_forward
* norm.cpp: remove try catch exception block
Sigbjørn Skjæret [Mon, 31 Mar 2025 09:07:07 +0000 (11:07 +0200)]
llava : fix clip loading GGUFs with missing description (#12660)