Search results

Eval bug: gemma output <unused24><unused24><unused24><unused24>

Status: Open.

#15 In AmesianX/TurboQuant;

· jarkevithwlad opened on Apr 5, 2026

Compile bug: warning: no usable GPU found Build When DGGML_HIP=ON

Status: Open.

#14 In AmesianX/TurboQuant;

· baramofme opened on Apr 4, 2026

Feature Request: Metal GPU Support for TurboQuant KV Cache Compression

Status: Open.

#12 In AmesianX/TurboQuant;

· richardokonicha opened on Apr 4, 2026

V-cache precision bug persists in v1.4.1 on head_dim=128 models (Qwen3-14b) — KV recall degradation from ctx ~1100t

Status: Open.

#11 In AmesianX/TurboQuant;

· primoco opened on Apr 4, 2026

Windows Support

Status: Open.

#8 In AmesianX/TurboQuant;

· Ezzz-dev opened on Apr 3, 2026

rocm build failed

Status: Open.

#5 In AmesianX/TurboQuant;

· dpblnt opened on Apr 2, 2026

pre-allocated tensor (cache_k_l3 (view)) in a buffer (Vulkan0) that cannot run the operation (SET_ROWS)

Status: Open.

#4 In AmesianX/TurboQuant;

· dpblnt opened on Apr 2, 2026