Search results

Eval bug: How to use Gemma 31B Qat with mtp?

Status: Open.

#64 In Anbeeld/beellama.cpp;

· ajcn02 opened on Jun 8, 2026

Eval bug: 0.3.2 compiled from latest source crashes

Status: Open.

#61 In Anbeeld/beellama.cpp;

· QBANIN opened on Jun 7, 2026

Eval bug: Sm tensor doesn't work

Status: Open.

#60 In Anbeeld/beellama.cpp;

· borisk1 opened on Jun 7, 2026

Eval bug: Vulkan/HIP is way too slow in comparison to llama.cpp

Status: Open.

#59 In Anbeeld/beellama.cpp;

· Ezzz-dev opened on Jun 6, 2026

Eval bug: DFlash kills inference speed on non-DFlash inference (MTP, tool calls)

Status: Open.

#58 In Anbeeld/beellama.cpp;

· ethernidee opened on Jun 6, 2026

Misc. bug: Unable to build for gfx906 on rocm 6.4

Status: Open.

#55 In Anbeeld/beellama.cpp;

· milpster opened on Jun 5, 2026

Eval bug: Vision CUDA crash (0.3.1 regression): Qwen3.6 35B A3B — missing context checkpoint invalidation on image token position jump

Status: Open.

#54 In Anbeeld/beellama.cpp;

· chromascope-x opened on Jun 5, 2026

Eval bug: dflash: invalid reduced-logits token -1 in draft

Status: Open.

#47 In Anbeeld/beellama.cpp;

· Riconec opened on May 31, 2026

Eval bug: suboptimal DFlash multi-slot performance

Status: Open.

#46 In Anbeeld/beellama.cpp;

· Anbeeld opened on May 31, 2026

Eval bug: Segfault at slot initialization with CUDA on SM75 (Turing) — ngl > 0

Status: Open.

#41 In Anbeeld/beellama.cpp;

· Bino5150 opened on May 26, 2026

profiling radeon 7900xt (20gb vram)

Status: Open.

#40 In Anbeeld/beellama.cpp;

· cryptopsy0 opened on May 26, 2026

Eval bug: low DFlash tps/AR for multi-GPU setups

Status: Open.

#39 In Anbeeld/beellama.cpp;

· Anbeeld opened on May 26, 2026