Search results

模型下载器 魔搭社区 页面经常崩溃

Status: Open.

#828 In jundot/omlx;

· 1369946045 opened on Apr 17, 2026

Failed to load Qwen3.6 DFlash

Status: Open.

#827 In jundot/omlx;

· cresseelia opened on Apr 17, 2026

Prefix cache hit causes Qwen 3.6 (hybrid linear attention) to lose tool calling ability

Status: Open.

#825 In jundot/omlx;

· Wysie opened on Apr 17, 2026

Failed to start 0.3.6 on macOS 15.7.4

Status: Open.

#821 In jundot/omlx;

· andy-wei opened on Apr 17, 2026

Failed to load the lmstudio-community/Qwen3.6-35B-A3B-MLX-6bit model

Status: Open.

#819 In jundot/omlx;

· VSavitar opened on Apr 17, 2026

[VLMBatchedEngine] Tool calling silently stops — model outputs text without tool_call on Qwen3.6-35B (works fine with BatchedEngine)

Status: Open.

#812 In jundot/omlx;

· Keenni opened on Apr 16, 2026

SIGABRT crash in BatchedEngine generation step — silent stop during tool calls (35B-A3B mxfp8, 9B oQ4)

Status: Open.

#811 In jundot/omlx;

· Keenni opened on Apr 16, 2026

Persistent-server throughput degrades under repeated structured-output chat completions on Apple Silicon

Status: Open.

#810 In jundot/omlx;

· jdiamand opened on Apr 16, 2026

TurboQuant setting doesn't seems to work?

Status: Open.

#808 In jundot/omlx;

· Xartorx opened on Apr 16, 2026

[Feature] Stop inference if the client closes the connection

Status: Open.

#804 In jundot/omlx;

· zulufoxtrot opened on Apr 16, 2026

v0.3.5 Reasoning Parser dropdown missing for Nemotron 3

Status: Open.

#801 In jundot/omlx;

· moritzchow opened on Apr 16, 2026

STT does not seem to work with MLX models

Status: Open.

#800 In jundot/omlx;

· fingon opened on Apr 16, 2026