Search results

Feature Request: Implement MotifForCausalLM

enhancementNew feature or requestNew feature or request

help wantedNeeds help from the communityNeeds help from the community

modelModel specificModel specific

porting

Status: Open.

#18055 In ggml-org/llama.cpp;

· pwilkin opened on Dec 15, 2025

Memory pool allocation bug prevents concurrent use of multiple LoRA adapters (GGML_ASSERT failure)

bugSomething isn't workingSomething isn't working

schedulerBugs related to the scheduler and memory allocator mechanismBugs related to the scheduler and memory allocator mechanism

Status: Open.

#18050 In ggml-org/llama.cpp;

· AntmanLFE opened on Dec 15, 2025

Compile bug: BUILD_TOOLS now requires BUILD_SERVER

bugSomething isn't workingSomething isn't working

buildCompilation issuesCompilation issues

Status: Open.

#18046 In ggml-org/llama.cpp;

· 65a opened on Dec 15, 2025

Misc. bug: Repeated command line options are discarded in router mode (--override-kv is broken)

bugSomething isn't workingSomething isn't working

regressionA regression introduced in a new build (something that was previously working correctly)A regression introduced in a new build (something that was previously working correctly)

server

Status: Open.

#18040 In ggml-org/llama.cpp;

· personalmountains opened on Dec 14, 2025

Misc. bug: Unhandled exception for non-existing --models-dir

bugSomething isn't workingSomething isn't working

low severityUsed to report low severity bugs in llama.cpp (e.g. cosmetic issues, non critical UI glitches)Used to report low severity bugs in llama.cpp (e.g. cosmetic issues, non critical UI glitches)

server

Status: Open.

#18037 In ggml-org/llama.cpp;

· personalmountains opened on Dec 14, 2025

Misc. bug: 2 GB at the start

bug-unconfirmed

needs evalNeeds maintainer evaluation for inclusion viabilityNeeds maintainer evaluation for inclusion viability

Status: Open.

#18024 In ggml-org/llama.cpp;

· tim-lebedkov opened on Dec 14, 2025

Feature Request: WebUI option to have lower CPU usage on web browser

enhancementNew feature or requestNew feature or request

server/webui

Status: Open.

#18023 In ggml-org/llama.cpp;

· mingodad opened on Dec 14, 2025

Compile bug: Container build fails due to incorrect LD_LIBRARY_PATH for DGX Spark/GB 10

bug-unconfirmed

Status: Open.

#18022 In ggml-org/llama.cpp;

· stelterlab opened on Dec 14, 2025

Feature Request: WebUI Conversations Export: add context and timestamp info

enhancementNew feature or requestNew feature or request

server/webui

Status: Open.

#18021 In ggml-org/llama.cpp;

· mingodad opened on Dec 14, 2025

Eval bug: ROCm model loading with Strix Halo (Ryzen 395) always dumps KV cache into shared memory

bug-unconfirmed

need feedbackTesting and feedback with results are neededTesting and feedback with results are needed

performanceSpeed related topicsSpeed related topics

RoCMIssues related to the RoCM backendIssues related to the RoCM backend

Status: Open.

#18011 In ggml-org/llama.cpp;

· Goldenkoron opened on Dec 13, 2025

Eval bug: Regression when running Devstral 2 123B with the latest llama.cpp version from Docker

bug-unconfirmed

need feedbackTesting and feedback with results are neededTesting and feedback with results are needed

Status: Open.

#18007 In ggml-org/llama.cpp;

· fergusq opened on Dec 13, 2025

Misc. bug: Since b7101 model output by bin-win-vulkan-x64 with quantized model and cache-type-k and cache-type-v is flawed

bugSomething isn't workingSomething isn't working

regressionA regression introduced in a new build (something that was previously working correctly)A regression introduced in a new build (something that was previously working correctly)

VulkanIssues specific to the Vulkan backendIssues specific to the Vulkan backend

Status: Open.

#17995 In ggml-org/llama.cpp;

· Zaakh opened on Dec 13, 2025