Issues

100% cpu spinning on tool call, generastion cannot be aborted

#2500

· schmorp opened

on Sep 26, 2026

[ Feature Request ] Update for Qwen image 2.1 support.

#2495

· Peterx2026 opened

on Sep 24, 2026

RX7900 GRE / Windows 11 / Gemma4 Model works in 1.119 (nocuda) but not in 1.121 (nocuda)

#2494

· 8u6man opened

on Sep 22, 2026

Prompt processing hangs indefinitely when prompt token count lands a few tokens past the batch/ubatch size (reproduced at batchsize=1024 → hang at ~1030 tokens, and batchsize=1536 → hang at ~1542 tokens)

#2493

· arnatoatadaink opened

on Sep 22, 2026

`--autofitpadding` for Secondary GPUs

#2492

· jobobby04 opened

on Sep 21, 2026

[Bug] `--routermode` + `--config` triggers unnecessary model reload on first API request

#2490

· VL4DST3R opened

on Sep 21, 2026

[Bug]: Visual artifacts appear on application window upon launch

#2489

· FckingChuck opened

on Sep 21, 2026

Kobold Lite fails to load when Web Audio is disabled in Firefox

#2485

· mokitoshi opened

on Sep 19, 2026

[Feature Request] Auto-save config to kcpps when launching

#2482

· Sydebrix opened

on Sep 19, 2026

OpenAI-compatible streaming: no `usage` chunk when the request contains `tools` (fake-stream path) — breaks client-side token accounting / auto-compaction

#2480

· 4eJIoBek1 opened

on Sep 18, 2026

[FIXED][Bug] Makefile LDFLAGS for LLAMA_VULKAN is shadowed by misplaced endif

#2476

· tncfdp opened

on Sep 17, 2026

[Feature Request] Please add a UI toggle for DP2A emulation (DISABLE_DP4A) to fix slow generation on NVIDIA CMP cards

#2472

· zotyfa-web opened

on Sep 15, 2026

Navigation Menu

All issues

100% cpu spinning on tool call, generastion cannot be aborted

[ Feature Request ] Update for Qwen image 2.1 support.

RX7900 GRE / Windows 11 / Gemma4 Model works in 1.119 (nocuda) but not in 1.121 (nocuda)

Prompt processing hangs indefinitely when prompt token count lands a few tokens past the batch/ubatch size (reproduced at batchsize=1024 → hang at ~1030 tokens, and batchsize=1536 → hang at ~1542 tokens)

`--autofitpadding` for Secondary GPUs

[Bug] `--routermode` + `--config` triggers unnecessary model reload on first API request

[Bug]: Visual artifacts appear on application window upon launch

Kobold Lite fails to load when Web Audio is disabled in Firefox

[Feature Request] Auto-save config to kcpps when launching

OpenAI-compatible streaming: no `usage` chunk when the request contains `tools` (fake-stream path) — breaks client-side token accounting / auto-compaction

[FIXED][Bug] Makefile LDFLAGS for LLAMA_VULKAN is shadowed by misplaced endif

[Feature Request] Please add a UI toggle for DP2A emulation (DISABLE_DP4A) to fix slow generation on NVIDIA CMP cards

Search results