* Add --override-tensors option to llama-bench
* Correct llama-bench --override-tensors to --override-tensor
* llama-bench: Update --override-tensors parsing to match --tensor-split, appear in test matrix.
* Make new llama-bench util functions static to fix Ubuntu CI
* llama-bench: Correct -ot corner cases (No -ot calls, leading and trailing empty -ot spans, etc.)
* fix wrong template in GLM4-0414
* fix spaces
* no bos token since it is already in the template
* moved the chatgml4 check to higher priority
* restored template for old GLM models
* moved the GLM4 template check in the correct place with correct check
* implment vision model architecture, gguf convertor
* handle window attention inputs
* add debug utils
* fix few incorrect tensor memory layout
* move position id remap out of ggml to avoid int32 cuda operations
* cleaning up
* ignore transformers Qwen2_5_xxx type check
* remove not so often use `qwen2vl-cli` debug functions
* remove commented-out code blocks
* fix attn weight scaling after rebase
* add `PROJECTOR_TYPE_QWEN2_5_VL`
* remove `KEY_USE_GLU_MLP`, `KEY_USE_RMS_NORM`
* replace `KEY_FULLATTN_BLK_IDX` with `KEY_WIN_ATTN_PATTERN`
* remove `attn_window_size` from gguf
* fix model conversion
* clean up
* fix merging problem
* add test
---------
Co-authored-by: Xuan Son Nguyen <son@huggingface.co>
* Force FP32 compute in cuBLAS GEMM
* Revert "Force FP32 compute in cuBLAS GEMM"
This reverts commit 6efd872732159ab88ee7b3c1d77ba5ebc83079bd.
* Force F32 compute in GLM4 ffn down
* Edit comment to clarify issue
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
RPC_CMD_SET_TENSOR always returns an empty response and we send this 4
times per token. We can improve TG speed if we don't wait for this empty
response.
The performance impact of this change depends on the network latency.
* cmake : do not include ./src as public for libllama
ggml-ci
* cmake : rework tests
ggml-ci
* llguidance : remove unicode include
ggml-ci
* cmake : make c++17 private
ggml-ci
* arg : clean up handling --mmproj with -hf
* rm change about no_mmproj
* Revert "rm change about no_mmproj"
This reverts commit 2cac8e0efb629d66c612f137e75d562f94bb9e6c.
* handle no_mmproj explicitly
* skip download mmproj on examples not using it
* tune matmul for gcn
* this one is more power efficient
* Update ggml/src/ggml-vulkan/ggml-vulkan.cpp
Co-authored-by: 0cc4m <picard12@live.de>
* disable this tune for the proprietary driver
---------
Co-authored-by: 0cc4m <picard12@live.de>
* add pixtral text model (vision is wip)
* cgraph ok, just missing 2D RoPE
* fix bad rebase
* first working version
* fix problem with img_break token
* support dynamic image size
* update docs
* update test script
* mtmd : merge `llava-cli` and `gemma3-cli` into single `mtmd-cli`
* support for minicpmv
* remove cpp files of llava and minicpmv
* update hot topics
* mtmd : add not supported msg for qwen2vl
* Update examples/llava/mtmd.cpp
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This restores the behavior from #491. This does not affect Ctrl+D's ability to
terminate --multiline-input lines (#1040).
This also actually implements #587: "If the user wants the text to end in a
newline, this should be accomplished by explicitly adding a newline by using
\ followed by return, then returning control by pressing return again."
Fixes#12949
* server : use std::move whenever possible
* use r-value ref
* Apply suggestions from code review
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* make task creation scoped
* restore std::move
* fix task_id not set correctly
* apply changes from suggestion
Co-authored-by: ggerganov <ggerganov@gmail.com>
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* mtmd : add more api around mtmd_image_tokens
* mtmd : ability to calc image hash
* shared_ptr for mtmd_image_tokens
* move hash to user-define ID (fixed)
* fix prompt_modified
* rm redundant data member
Add RPC_CMD_HELLO for getting the version of the protocol implemend by
the server. Follow the semantic versioning rules at https://semver.org
Hopefully this bring better user experience when we make breaking
changes at the protocol level and avoid issues like #12465
* graph : make mla compatible with FA
* metal : add exp FA kernels for DeepSeek models
ggml-ci
* llama : minor naming updates
ggml-ci
* ggml : disable FA for DS head sizes
* tests : add FA tests for MLA shapes
ggml-ci
Submit operators using asynchronous threads to improve performance.
Use the environment variable GGML_CANN_ASYNC_MODE to control whether
asynchronous submission is enabled. It is disabled by default.
Testing shows a 10%–20% performance improvement in scenarios with
small parameter sizes, especially in quantized models.
The Granite's FIM tokens are very similar to Qwen's; it's just that
they use underscore instead of a dash. So <fim_middle> for example
instead of <fim-middle>.
Opening up tokenizer_config.json in ibm-granite/granite-3.3-8b-base
shows:
```
"<fim_prefix>",
"<fim_middle>",
"<fim_suffix>",
"<fim_pad>",
...
"<reponame>",
```
The grouped query attention optmization doesn't require a power of two ratio,
the only thing relying on it was the modulo operation written as bitwise &.
split_k need not depend on gqa_ratio - enable it any time there's only one
workgroup in the X dimension. The shader gets the split index from the x coord,
and multiple workgroups in the X dimension (pre-split) indicates a larger
FA operation that wouldn't need splitting.
* opencl: refactor - split the kernel files
---------
Co-authored-by: Shangqing Gu <quic_shawngu@quicinc.com>
* opencl: split more kernels into separate files
* opencl: specify subgroup size instead of querying it
* opencl: refine Adreno cl compiler version parsing
* opencl: skip some kernels not used by Adreno on old compilers
* opencl: refine logic for selecting Adreno kernels
* opencl: refine Adreno cl compiler version
* opencl: cleanup preprocessor for kernels
* opencl: consider Adreno CL compiler on Windows
* opencl: add final newline for `mul_mv_f16_f16.cl`
---------
Co-authored-by: Shangqing Gu <quic_shawngu@quicinc.com>
Replace compile-time `GGML_HIP_UMA` with environment variable `GGML_CUDA_ENABLE_UNIFIED_MEMORY`. This unifies the usage on NVIDIA and AMD GPUs, and allows a single binary to be shared between integrated and dedicated GPUs.
* Merged using squash to remove all noise commit messages
* Force flash attention off for `LLM_ARCH_DEEPSEEK2` - embedding too large
* Removed 3 conts (2x RoPE and 1x RMS-norm)
* Changed to use `<cmath>` instead of `<math.h>`
* Reverted removal of the 3 conts
* Used `reshape` in `llm_graph_context::build_attn_mha()`
* Use `k_pe = ggml_reshape`
* Removed the 3 conts again
* Removed the 3D views of `wk_b` and `wv_b`, and just save and 3D in GGUF
* Removed MQA optimisation from `build_attn_mha()` as no gains now
* Simplified `is_mla` branch in `llm_build_deepseek2()`
* Removed `build_attn_mla` and added `nullptr` to all `build_atnn` calls
* Fixed call to `build_attn` in `llm_build_t5_enc`
Multiple optional memory pools are provided for CANN, including VMM,
priority queue-based, and traditional memory pools.
1.When the memory pool is available and GGML_CANN_DISABLE_VMM_POOL
is not defined, the VMM pool is selected by default.
2.Otherwise, if GGML_CANN_ENABLE_BUF_PRIO_POOL is defined,
the priority queue-based memory pool is used.
3.If neither condition is met, the default memory pool is used.
* Add llama_model_quantize_params parameters
* Add new quantize parameters parsing and validation
* Update usage
* Add new parameters defaults
* Add new quantization parameters logic
* Add llama_model_quantize_params parameters
* Add new quantize parameters parsing and validation
* Update usage
* Add new parameters defaults
* Add new quantization parameters logic
* Minor refactoring as per the contributors' coding guidelines
* Update descriptions to match existing style
* Add llama_model_quantize_params parameters
* Add new quantize parameters parsing and validation
* Update usage
* Add new parameters defaults
* Add new quantization parameters logic
* Minor refactoring as per the contributors' guidelines
* Implement general --tensor-type instead of tensor-specific command option
* Fix implied type bug
* Restore missing #includes
* Add regex capability for tensor selection
* Refactor function name and update ALLOWED_TENSOR_TYPE
* Add missing #include
* Handle edge case when tensor name is cls.output
* Minor logging improvement
The current usage of the SYCL-Graph extension checks for
the `sycl_ext_oneapi_graph` device aspect. However, it is also
possible to support `sycl_ext_oneapi_limied_graph` devices that
don't support update
* SYCL: Add fp16 support to some elementwise OP kernels
* remove comment
ggml-ci
* Use static_cast directly
* remove not needed cast from tanh
* Use static cast and remove unneeded castings
* Adjust device_support_op for unary OPs
* Use cast_data and typed_data struct to deduplicate casting code
This commit adds a check for the visionos build version used with vtool
in build-xcframework.sh. The script now checks the Xcode version and
determines whether to use "xros" or "visionos" for the build version.
This commit also uses xcrun for the vtool so that the version of vtool
in xcode command line tools is used instead of the one in the system
path.
Refs: https://github.com/ggml-org/whisper.cpp/pull/2994#issuecomment-2773292223
Johannes Gäßler