32 Releases published by 1 person

• b4800

  published Mar 2, 2025

• b4801

  published Mar 3, 2025

• b4803

  published Mar 3, 2025

• b4804

  published Mar 3, 2025

• b4805

  published Mar 3, 2025

• b4806

  published Mar 3, 2025

• b4818

  published Mar 3, 2025

• b4819

  published Mar 3, 2025

• b4820

  published Mar 4, 2025

• b4821

  published Mar 4, 2025

• b4823

  published Mar 4, 2025

• b4824

  published Mar 5, 2025

• b4826

  published Mar 5, 2025

• b4827

  published Mar 5, 2025

• b4829

  published Mar 5, 2025

• b4830

  published Mar 5, 2025

• b4831

  published Mar 5, 2025

• b4832

  published Mar 6, 2025

• b4833

  published Mar 6, 2025

• b4834

  published Mar 6, 2025

• b4835

  published Mar 6, 2025

• b4836

  published Mar 6, 2025

• b4837

  published Mar 6, 2025

• b4846

  published Mar 7, 2025

• b4847

  published Mar 7, 2025

• b4848

  published Mar 7, 2025

• b4849

  published Mar 7, 2025

• b4851

  published Mar 7, 2025

• b4853

  published Mar 7, 2025

• b4854

  published Mar 7, 2025

• b4855

  published Mar 7, 2025

• b4856

  published Mar 8, 2025

47 Pull requests merged by 26 people

• authors : update

  #12271 merged Mar 8, 2025

• ggml-backend : make path_str compatible with C++20

  #12269 merged Mar 8, 2025

• server : infill gen ends on new line

  #12254 merged Mar 7, 2025

• ggml : skip intermediate .air file when compiling .metallib

  #12247 merged Mar 7, 2025

• sync : ggml

  #12248 merged Mar 7, 2025

• ggml-cpu: faster AVX2 variant for IQ1_M

  #12216 merged Mar 7, 2025

• ci : fix save-load test invokations

  #12245 merged Mar 7, 2025

• server: log original chat template parsing error

  #12233 merged Mar 7, 2025

• sync: minja (support QwQ-32B)

  #12235 merged Mar 7, 2025

• metal : simplify kernel arguments using a struct (#3229)

  #12194 merged Mar 7, 2025

• Fix HIP rocWMMA CI build break

  #12230 merged Mar 7, 2025

• metal : fix default.metallib build

  #12224 merged Mar 7, 2025

• opencl: Noncontiguous norm, rms_norm, disable fp16 for some ops

  #12217 merged Mar 7, 2025

• cmake : fix undefined reference errors for std::filesystem in ggml (#12092)

  #12094 merged Mar 6, 2025

• Update README.md

  #12229 merged Mar 6, 2025

• CUDA: fix FA logic for PTX 7.0 and CC >= 7.5

  #12222 merged Mar 6, 2025

• HIP: rocWMMA documentation and enabling in workflow builds

  #12179 merged Mar 6, 2025

• docs: update function-calling.md w/ template override needed by functionary-small-v3.2

  #12214 merged Mar 6, 2025

• llava: add big-endian conversion for image encoder

  #12218 merged Mar 6, 2025

• HIP/CUDA: set the paramerter value in maintain_cuda_graph instead of replaceing it.

  #12209 merged Mar 6, 2025

• android : Calculate required KV cache size by summing up tokens size and response token length (#12211)

  #12212 merged Mar 6, 2025

• opencl: Fix not enough space in the buffer

  #12197 merged Mar 6, 2025

• opencl: Fix ulong kernel args were set from int variables

  #12174 merged Mar 6, 2025

• opencl:Fix profile-related errors

  #12095 merged Mar 6, 2025

• ggml-cpu: Faster IQ1 mul_mat_vec on AVX2 using BMI2 instructions

  #12154 merged Mar 6, 2025

• SYCL: Disable f16 Unary OPs as not supported by the kernels

  #12201 merged Mar 5, 2025

• ggml : refactor metal library loading to avoid GGMLMetalClass ODR

  #12200 merged Mar 5, 2025

• ci : add fetch-depth to xcframework upload

  #12195 merged Mar 5, 2025

• tool-call: fix Qwen 2.5 Coder support, add micro benchmarks, support trigger patterns for lazy grammars

  #12034 merged Mar 5, 2025

• ci : fix xcframework artifact tag

  #12191 merged Mar 5, 2025

• ci : remove xframework upload

  #12190 merged Mar 5, 2025

• Server: Cache position calculation error(#12160)

  #12161 merged Mar 5, 2025

• llama : add xcframework build script

  #11996 merged Mar 5, 2025

• Some portability improvements from trying to build with Visual Studio 2017

  #12150 merged Mar 4, 2025

• readme : fix roadmap link

  #12185 merged Mar 4, 2025

• main: allow preloading conversation with -p and add -st / --single-turn

  #12145 merged Mar 4, 2025

• server: fix response_format w/ json_schema.schema

  #12168 merged Mar 4, 2025

• Add GGML_HIP_ROCWMMA_FATTN to enable rocWMMA for FlashAttention

  #12032 merged Mar 3, 2025

• sync : ggml

  #12104 merged Mar 3, 2025

• ci : set GITHUB_ACTIONS to true for server tests

  #12162 merged Mar 3, 2025

• tts: add speaker file support

  #12048 merged Mar 3, 2025

• test-backend-ops : add option -p to filter by op params

  #12155 merged Mar 3, 2025

• Fix kleidiai build

  #12159 merged Mar 3, 2025

• Adding UTF-8 support to linenoise.cpp

  #12111 merged Mar 3, 2025

• webui : add ?m=... and ?q=... params

  #12148 merged Mar 3, 2025

• SYCL: Move CPY kernels to a separate file and add few missing kernels

  #12133 merged Mar 3, 2025

• ggml-backend : keep paths in native string type when possible

  #12144 merged Mar 2, 2025

28 Pull requests opened by 24 people

• build: fix build error when build source code on Windows

  #12157 opened Mar 3, 2025

• CUDA/HIP: refractor mmqv to unify the calculation of nwarps and rows per block between host and device code.

  #12177 opened Mar 4, 2025

• llama : refactor llama_context, llama_kv_cache, llm_build_context (v2)

  #12181 opened Mar 4, 2025

• CUDA: Improve flash decoding kernel GPU occupancy for BS=1 case

  #12183 opened Mar 4, 2025

• fix: AVX2 intrinsics, const correctness, and SIMD headers

  #12186 opened Mar 4, 2025

• vulkan: double buffer scale caches

  #12188 opened Mar 4, 2025

• libfuse3 supported mounting split gguf's to a single in-memory file

  #12189 opened Mar 5, 2025

• SYCL: Rename oneMKL to oneMath

  #12192 opened Mar 5, 2025

• feat(CMakeLists): Add MSVC-specific compiler warning flags in CMake configuration

  #12206 opened Mar 5, 2025

• build: build llama.cpp + ggml-qnn in pure command line mode on x86-64 Windows

  #12215 opened Mar 6, 2025

• opencl: use OpenCL C standard supported by the device

  #12221 opened Mar 6, 2025

• Optimized DeepSeek V2/V3 implementation (MLA + flash attention)

  #12227 opened Mar 6, 2025

• tests: use adaptive number of threads

  #12236 opened Mar 6, 2025

• Issues while enabling MMA support on AIX machines

  #12241 opened Mar 7, 2025

• Fix rocWMMA build documentation

  #12243 opened Mar 7, 2025

• clang-tidy : disable bugprone-branch-clone

  #12244 opened Mar 7, 2025

• server : Add verbose output to OAI compatible chat endpoint.

  #12246 opened Mar 7, 2025

• main : add -sysf / --system-prompt-file (#12249)

  #12250 opened Mar 7, 2025

• vulkan: Adjust coopmat2 tile sizes and selection heuristic

  #12258 opened Mar 7, 2025

• vulkan: optimization proposals for coopmat1 mul_mm

  #12260 opened Mar 7, 2025

• Add simple-tts example

  #12261 opened Mar 8, 2025

• doc: add text-based diagram of software architecture in toplevel README.md

  #12263 opened Mar 8, 2025

• metal: Cache compiled library at device level

  #12265 opened Mar 8, 2025

• vulkan: fix coopmat shader generation when cross-compiling

  #12272 opened Mar 8, 2025

• vulkan: Pad N dimension of B matrix for coopmat2 perf, to avoid bounds checking

  #12273 opened Mar 8, 2025

• (research) experiment with phi-4-multimodal vision support

  #12274 opened Mar 8, 2025

• Refactoring '-o' option

  #12278 opened Mar 9, 2025

• server: fix "--grammar-file" parameter

  #12285 opened Mar 9, 2025

29 Issues closed by 18 people

• Misc. bug: GPU Support Missing in Version >=0.3.5 on Windows with CUDA 12.4 and RTX 3090

  #12283 closed Mar 9, 2025

• Compile bug: Unable to make the ./main folder and run fine tuned model.

  #12257 closed Mar 7, 2025

• Support for AMD iGPU?

  #12239 closed Mar 7, 2025

• llama : add test for saving/loading sessions to the CI

  #2631 closed Mar 7, 2025

• Eval bug: Jinja parser not working with QwQ-32B

  #12231 closed Mar 7, 2025

• Misc. bug: llama.swiftui simulator error

  #12219 closed Mar 7, 2025

• metal : simplify kernel arguments using a struct

  #3229 closed Mar 7, 2025

• Eval bug: llama.cpp returns gibberish on Intel Core Ultra 7 (155H) with ARC iGPU

  #12096 closed Mar 6, 2025

• [Solved]Model generation speed significantly slows down when using MiroStat V2

  #12220 closed Mar 6, 2025

• Misc. bug: llama-cli's inference result seems not correct on 64bit-Windows

  #12226 closed Mar 6, 2025

• Eval bug: Granite Vision 3.1 and 3.2 Surgery Script Found 0 Tensors to Extract

  #12202 closed Mar 6, 2025

• Eval bug: Incorrect KV cache calculation in llama.android example

  #12211 closed Mar 6, 2025

• Compile bug: RISCV compilation HELP

  #12170 closed Mar 6, 2025

• Misc. bug: SYCL out of memory error

  #11044 closed Mar 6, 2025

• EoS Tokenization issue for Nemo 12b

  #11299 closed Mar 6, 2025

• Eval bug: bartowski/functionary-small-v3.2-GGUF:Q4_K_M model prepends "assistant\n" to text responses when tools are provided

  #12213 closed Mar 6, 2025

• Misc. bug: Q4_0 repacking results in double RAM usage

  #12149 closed Mar 5, 2025

• Misc. bug: Corrupted HF models

  #12207 closed Mar 5, 2025

• Misc. bug: `json_schema` under `response_format` is not working on OpenAI compatible API endpoint `v1/chat/completions`

  #11988 closed Mar 4, 2025

• Feature Request: Split model over multiple Vulkan GPUs

  #11004 closed Mar 4, 2025

• Misc. bug: strange reducing memsize type to 32bit without dev comment

  #11293 closed Mar 4, 2025

• Bug: Flash Attention performs worse under ROCM

  #10439 closed Mar 3, 2025

• Compile bug: How to build llama.android example with -DGGML_VULKAN=ON through android studio.

  #12085 closed Mar 3, 2025

• Feature Request: Avoid xml use in tool call instructions.

  #12153 closed Mar 3, 2025

• Eval bug: llama cpp becomes slower as the number of threads -t increases

  #11247 closed Mar 3, 2025

• qwen model quantized with AWQ and lora weights

  #11277 closed Mar 3, 2025

• Misc. bug: --file flag not working

  #12138 closed Mar 2, 2025

• Eval bug: Crash with filesystem error when run while in a directory containing files with certain names

  #11198 closed Mar 2, 2025

• Misc. bug: hipGraph causes a crash in hipGraphDestroy

  #11949 closed Mar 2, 2025

41 Issues opened by 37 people

• Compile bug: 执行python3 convert-hf-to-gguf.py D:\DevelopSoftware\Ollamamodel\DeepSeek-R1-Medical-COT-500命令，没有响应

  #12286 opened Mar 9, 2025

• Misc. bug: An error occurred when committing using the pre-commit config in the project

  #12284 opened Mar 9, 2025

• Eval bug: BLAS backend crash in ggml_backend_dev_backend_reg

  #12282 opened Mar 9, 2025

• Compile bug: undefined reference to `ggml_set_f32_nd'

  #12281 opened Mar 9, 2025

• Misc. bug: tool call issues with hf unsloth/Qwen2.5-Coder-7B-Instruct-128K-GGUF

  #12279 opened Mar 9, 2025

• Eval bug: GPU Hang Error on Metal backend

  #12277 opened Mar 8, 2025

• Feature Request: grammar / json schema with reasoning format. Allow model free to think but strict to answer.

  #12276 opened Mar 8, 2025

• Misc. bug: QwQ 32B doesn't put the reasoning content in `message.reasoning_content`

  #12275 opened Mar 8, 2025

• Feature Request: Add support for InstellaForCausalLM model architecture

  #12270 opened Mar 8, 2025

• Misc. bug: Misc. bug: cannot convert GLM-4v-9B- (glm-4v-9b) to GGUF format #11263

  #12266 opened Mar 8, 2025

• Eval bug: server API endpoint not respecting `n_predict` with `-2` (until context filled)

  #12264 opened Mar 8, 2025

• Eval bug: Segfault at the end of the cache (cache defragmentation?)

  #12259 opened Mar 7, 2025

• Eval bug: Llama server <tool_call> is occasionally not parsed as json, and is in content rather than tool_calls

  #12256 opened Mar 7, 2025

• Misc. bug: convert_hf_to_gguf failing for deepseek-r1 full

  #12255 opened Mar 7, 2025

• Eval bug: garbage output right after kv-cache defragmentation for CPU backend

  #12253 opened Mar 7, 2025

• Misc. bug: llama-server: SegFault with json_schema containing unsupported pattern

  #12252 opened Mar 7, 2025

• Eval bug: QWQ generates repeated text when running with reduced context length

  #12251 opened Mar 7, 2025

• Misc. bug:

  #12249 opened Mar 7, 2025

• Feature Request: Convert deepseek-v3's mtp module to gguf and quantize to q4km

  #12242 opened Mar 7, 2025

• Issues while enabling MMA support on AIX machines

  #12240 opened Mar 7, 2025

• FA bug causes Memory access fault by GPU

  #12238 opened Mar 7, 2025

• Prompt eval is 5x slower than in Ollama and maxes out the CPU

  #12237 opened Mar 6, 2025

• Eval bug: Excessive stack usage during tool calling

  #12234 opened Mar 6, 2025

• Eval bug: Phi-4 mini in iOS with xcframework

  #12232 opened Mar 6, 2025

• Eval bug: Failed to Load Granite Vision on s390x (llama-llava-cli)

  #12225 opened Mar 6, 2025

• Misc. bug: ALL gguf models fail to run (no log, docker exit code 139),

  #12205 opened Mar 5, 2025

• Misc. bug: llama-rpc crashes when deciding memory on CPU with CUDA_VISIBLE_DEVICES=""

  #12203 opened Mar 5, 2025

• [Metal] Context init optimization opportunity: metal library is compiled for every llama context

  #12199 opened Mar 5, 2025

• Eval bug: Crash on lazy grammar

  #12196 opened Mar 5, 2025

• CUDA: Improve flash decoding kernel occupancy for BS=1 case

  #12182 opened Mar 4, 2025

• Misc. bug: [Server] Crashes with a coredump during termination

  #12180 opened Mar 4, 2025

• Misc. bug: Denial of Service (crash) when using verbose output with input tokens that are not in printable range.

  #12178 opened Mar 4, 2025

• Eval bug: Server returns 500 error on /api/generate and /api/chat requests

  #12176 opened Mar 4, 2025

• Misc. bug: HIP compilation together with -DGGML_CPU_ALL_VARIANTS=ON does not load the model or detects the GPU

  #12175 opened Mar 4, 2025

• Misc. bug: The inference speed of llama-server is one-third of that of llama-cli

  #12171 opened Mar 4, 2025

• Compile bug: issue compiling in ubuntu (desktop and server version) using virtualbox

  #12164 opened Mar 3, 2025

• Eval bug: loading model: vk::PhysicalDevice::createDevice: ErrorExtensionNotPresent. Not falling back to CPU

  #12163 opened Mar 3, 2025

• Misc. bug: Calculating the position of kv cache error in llama sever

  #12160 opened Mar 3, 2025

• Eval bug: The answers have some problems with the example/llama.android

  #12158 opened Mar 3, 2025

• CUDA: HIP: maintain_cuda_graph use of cudaGraphKernelNodeGetParams is incorrect.

  #12152 opened Mar 2, 2025

• Replacement for deprecated codevct string conversion

  #12151 opened Mar 2, 2025

74 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• llama-tts : add -o option

  #12042 commented on Mar 8, 2025 • 12 new comments

• tool-call: add support for tool-calls using Model Context Protocol

  #11556 commented on Mar 9, 2025 • 4 new comments

• Vulkan: Add DP4A MMQ and Q8_1 quantization shader

  #12135 commented on Mar 4, 2025 • 1 new comment

• Eval bug: <think> tag with DeepSeek-R1-Distill-Qwen-1.5B-Q5_K_M.gguf

  #11325 commented on Mar 9, 2025 • 0 new comments

• Feature Request: allow to run on CPU despite backend initialization failure.

  #11584 commented on Mar 9, 2025 • 0 new comments

• Feature Request: allow setting jinja chat template from server webui

  #11689 commented on Mar 9, 2025 • 0 new comments

• Eval bug: Inconsistent <think> Tag Output in simple-chat vs. llama-cli with DeepSeek-R1-Distill-Qwen-7B-Q4_K_M Model

  #11702 commented on Mar 9, 2025 • 0 new comments

• Misc. bug: The test-chat fails with std::runtime_error

  #11705 commented on Mar 9, 2025 • 0 new comments

• Feature Request: (webui) read data from /props endpoint and use it on the webui

  #11717 commented on Mar 9, 2025 • 0 new comments

• Feature Request: (webui) add import / export function for ALL conversations

  #11718 commented on Mar 9, 2025 • 0 new comments

• Misc. bug: add tool_calls id in response in server

  #11992 commented on Mar 8, 2025 • 0 new comments

• Feature Request: Support for Phi4MMForCausalLM Architecture

  #12117 commented on Mar 8, 2025 • 0 new comments

• Misc. bug: Missing <think> tag in response (DeepSeek R1)

  #11861 commented on Mar 8, 2025 • 0 new comments

• ggml : add ANE backend

  #10453 commented on Mar 8, 2025 • 0 new comments

• Misc. bug: vulkan on 6900xt

  #12147 commented on Mar 8, 2025 • 0 new comments

• Regarding llama-bench and llama-parallel commands

  #12106 commented on Mar 8, 2025 • 0 new comments

• Misc. bug: ggml files conflict between llama.cpp and whisper.cpp

  #11303 commented on Mar 8, 2025 • 0 new comments

• Compile bug: Vulkan can not work on Android (cross-compilation from linux) - Aborted without explaination

  #11327 commented on Mar 8, 2025 • 0 new comments

• Eval bug: using rpc，report error [Inferior 1 (process 290070) detached]

  #11431 commented on Mar 8, 2025 • 0 new comments

• Eval bug: Segmentation fault on image encoder quantization

  #11683 commented on Mar 8, 2025 • 0 new comments

• sycl: cleanup oneDNN related code

  #12097 commented on Mar 4, 2025 • 0 new comments

• vulkan: subgroup size test

  #12087 commented on Mar 8, 2025 • 0 new comments

• [WIP]backend: Integrating QNN (Qualcomm AI Engine Direct) as a dedicated backend for Qualcomm NPUs

  #12063 commented on Mar 8, 2025 • 0 new comments

• ggml-cpu: add arm64 CPU feature check for OpenBSD, FreeBSD

  #11939 commented on Mar 5, 2025 • 0 new comments

• llama : private llama_batch

  #11875 commented on Mar 7, 2025 • 0 new comments

• Add supports for Janus vision encoder and projector [WIP]

  #11646 commented on Mar 7, 2025 • 0 new comments

• Add support for Deepseek-R1 flash attention

  #11557 commented on Mar 7, 2025 • 0 new comments

• Optimized DeepSeek V2/V3 implementation (MLA)

  #11446 commented on Mar 6, 2025 • 0 new comments

• llama : add option to override model tensor buffers

  #11397 commented on Mar 6, 2025 • 0 new comments

• llama : second attempt to refactor vision API

  #11292 commented on Mar 9, 2025 • 0 new comments

• SYCL: Fixes for building SYCL backend for AMD GPUs

  #10851 commented on Mar 6, 2025 • 0 new comments

• Introduce Graph Profiler

  #9659 commented on Mar 6, 2025 • 0 new comments

• Changes for the existing quant strategies / FTYPEs and new ones

  #8836 commented on Mar 8, 2025 • 0 new comments

• Rebalancing Metal threads workload in dot product kernel kernel_mul_mv_f16_f32_l4

  #7522 commented on Mar 8, 2025 • 0 new comments

• support MiniCPM-V-2

  #6919 commented on Mar 7, 2025 • 0 new comments

• Feature Request: Enable cuda 11.4 and cuda arch 3.7

  #12140 commented on Mar 9, 2025 • 0 new comments

• Research: Performance differences between Metal (macOS) and Vulkan (Linux)

  #10982 commented on Mar 9, 2025 • 0 new comments

• Compile bug: Nix + cross compilation + Vulkan doesn't work

  #11654 commented on Mar 8, 2025 • 0 new comments

• Feature Request: YuE (music gen)

  #11467 commented on Mar 5, 2025 • 0 new comments

• Misc. bug: Converting sftensor model

  #11497 commented on Mar 5, 2025 • 0 new comments

• Feature Request: resize an existing context

  #11577 commented on Mar 5, 2025 • 0 new comments

• Llama-3.2 11B Vision Support

  #9643 commented on Mar 4, 2025 • 0 new comments

• Misc. bug: llama-server throws "Unsupported param: tools"

  #10920 commented on Mar 4, 2025 • 0 new comments

• Misc. bug: model warmup doesn't work correctly for MoE models

  #11163 commented on Mar 4, 2025 • 0 new comments

• Eval bug: -sm row performance on NVidia multy-gpu config is extremely low on the long contexts after b3990

  #11510 commented on Mar 4, 2025 • 0 new comments

• Feature Request: add --no-warmup to llama-qwen2vl-cli

  #11526 commented on Mar 4, 2025 • 0 new comments

• Misc. bug: File "train-text-from-scratch" missing

  #11561 commented on Mar 4, 2025 • 0 new comments

• Flash attention implementations do not handle case where value vectors have different dimension from query vectors

  #7343 commented on Mar 3, 2025 • 0 new comments

• changelog : `libllama` API

  #9289 commented on Mar 3, 2025 • 0 new comments

• changelog : `llama-server` REST API

  #9291 commented on Mar 3, 2025 • 0 new comments

• Feature Request: mixed ROCm+CUDA possible?

  #11506 commented on Mar 3, 2025 • 0 new comments

• Misc. bug: server api endpoint /completion ignoring grammar parameter

  #11544 commented on Mar 3, 2025 • 0 new comments

• Compile bug: _WIN32_WINNT not, or not correctly set when compiling with clang and "MinGW Makefiles" generator

  #11542 commented on Mar 3, 2025 • 0 new comments

• Misc. bug: error when converting lora to gguf (ERROR:lora-to-gguf:Unexpected name 'base_model.model.lm_head.weight': Not a lora_A or lora_B tensor)

  #11554 commented on Mar 3, 2025 • 0 new comments

• Misc. bug: Quantization of deepseek r1 qwen models fails when using K quants.

  #11560 commented on Mar 3, 2025 • 0 new comments

• Misc. bug: vulkan on Adreno GPU

  #12139 commented on Mar 2, 2025 • 0 new comments

• Misc. bug: Sporadic MUL_MAT Failures in test-backend-ops for Nvidia backend

  #11972 commented on Mar 7, 2025 • 0 new comments

• Feature Request: Support for Deepseek Janus-Pro-7B & Janus-1.3B

  #11490 commented on Mar 7, 2025 • 0 new comments

• Feature Request: Qwen 2.5 VL

  #11483 commented on Mar 7, 2025 • 0 new comments

• Eval bug: ~~Q2_K and Q3_K~~ Q8_0 not working on Vulkan anymore on RX 5700XT

  #10710 commented on Mar 7, 2025 • 0 new comments

• Misc. bug: AMD Rcom command error only with cli tools

  #11509 commented on Mar 7, 2025 • 0 new comments

• Compile bug: ARMv7 NEON FP16 Intrinsic Errors When Cross-Compiling with Android NDK r26b

  #11636 commented on Mar 7, 2025 • 0 new comments

• Eval bug: qwen2-vl failed to process while using the HIP in windows 11

  #11638 commented on Mar 7, 2025 • 0 new comments

• Eval bug: CANNOT LINK EXECUTABLE "./llama-cli": library "libomp.so" not found: needed by main executable

  #11979 commented on Mar 7, 2025 • 0 new comments

• Feature Request: Proposing User-Customizable RAG Integration in llama.cpp: A Path to Enhanced Contextual Retrieval

  #12129 commented on Mar 6, 2025 • 0 new comments

• Eval bug: getting assertion error when trying to use a gguf quantized model at inference "GGML_ASSERT(n_outputs_enc > 0 && "call llama_encode() first") failed"

  #12080 commented on Mar 6, 2025 • 0 new comments

• Compile bug: llama.cpp-master/ggml/src/ggml-cpu/ggml-cpu-aarch64.cpp:80:54:error '_mm256_set_m128i' was not declared in this scope

  #11385 commented on Mar 6, 2025 • 0 new comments

• Misc. bug: Vulkan Q4_K_M inference speed degradation

  #11559 commented on Mar 6, 2025 • 0 new comments

• Feature Request: Ship llama.cpp binaries in AppImage format

  #11579 commented on Mar 6, 2025 • 0 new comments

• Compile bug: Not compilable with MACOSX_DEPLOYMENT_TARGET < 10.15

  #11612 commented on Mar 6, 2025 • 0 new comments

• Misc. bug: CTRL-ENTER not visibly shown in web UI prompt

  #11586 commented on Mar 6, 2025 • 0 new comments

• Misc. bug: Webui: Default light theme code blocks are not visible

  #11623 commented on Mar 6, 2025 • 0 new comments

• Refactor: llama-impl.h should be private

  #11630 commented on Mar 6, 2025 • 0 new comments

• Compile bug: undefined reference to std::filesystem

  #10978 commented on Mar 5, 2025 • 0 new comments