Search results

Feature Request: DeepSeek-V3.2-Exp support?

enhancementNew feature or requestNew feature or request

Status: Open.

#834 In ikawrakow/ik_llama.cpp;

· kkontosis opened on Oct 15, 2025

Bug: Crash in FlashQKV<128, 8, 128>::normalize_and_store_1row failing with GLM 4.5, AVX2, -amb/-ub/-b

Status: Open.

#809 In ikawrakow/ik_llama.cpp;

· os360 opened on Sep 29, 2025

Feature Request: Add <code>--device</code> and <code>--device-draft</code> parameters

enhancementNew feature or requestNew feature or request

Status: Open.

#785 In ikawrakow/ik_llama.cpp;

· MRGRD56 opened on Sep 18, 2025

Bug: PP Perf Regression after CUDA Graphs PR#689

Status: Open.

#765 In ikawrakow/ik_llama.cpp;

· usrlocalben opened on Sep 7, 2025

Bug: attention max batch option doesn't work for GLM 4.5

Status: Open.

#755 In ikawrakow/ik_llama.cpp;

· ghost opened on Sep 3, 2025

Bug: Partial parse: incomplete tool call

Status: Open.

#750 In ikawrakow/ik_llama.cpp;

· magikRUKKOLA opened on Sep 2, 2025

Bug: (Speculative decoding) Massive slowdown when going past draft model's ctx size (when -cd &lt; -c)

Status: Open.

#732 In ikawrakow/ik_llama.cpp;

· alint77 opened on Aug 26, 2025

Bug: (Speculative decoding) llama-server hangs but doesn't crash when draft's KV allocation fails (draft KV OOM)

Status: Open.

#731 In ikawrakow/ik_llama.cpp;

· alint77 opened on Aug 26, 2025

Feature Request: add support for vision model InternVL3_5

enhancementNew feature or requestNew feature or request

Status: Open.

#730 In ikawrakow/ik_llama.cpp;

· kiron111 opened on Aug 26, 2025

Bug: Surprising CUDA OOM during PP

wontfixThis will not be worked onThis will not be worked on

Status: Open.

#687 In ikawrakow/ik_llama.cpp;

· usrlocalben opened on Aug 12, 2025

Feature Request: openPangu-Ultra-MoE-718B support

enhancementNew feature or requestNew feature or request

help wantedExtra attention is neededExtra attention is needed

Status: Open.

#671 In ikawrakow/ik_llama.cpp;

· Lissanro opened on Aug 5, 2025

Feature Request: Multi NUMA Tensor Parallel

enhancementNew feature or requestNew feature or request

Status: Open.

#663 In ikawrakow/ik_llama.cpp;

· aikitoria opened on Jul 31, 2025