Search results

set min_length=1, remove audio less than 150ms. following errors stil stop training

Status: Open.

#130 In QwenLM/Qwen3-ASR;

· housebaby opened on Mar 19, 2026

Endless repetition in output

Status: Open.

#129 In QwenLM/Qwen3-ASR;

· andypotato opened on Mar 19, 2026

Evaluation procedure？

Status: Open.

#128 In QwenLM/Qwen3-ASR;

· vincentLehighYe opened on Mar 18, 2026

Using VLLM for backend inference in qwen3-asr may cause it to freeze (GPU memory is not released and utilization is 0%).

Status: Open.

#126 In QwenLM/Qwen3-ASR;

· Lee-xeo opened on Mar 16, 2026

不再适配 vllm 0.16.0+

Status: Open.

#119 In QwenLM/Qwen3-ASR;

· Lee-xeo opened on Mar 9, 2026

slow inference after finetuning

Status: Open.

#118 In QwenLM/Qwen3-ASR;

· binnybn98 opened on Mar 9, 2026

Why can't vllm be found after installation?

Status: Open.

#110 In QwenLM/Qwen3-ASR;

· SunYong0821 opened on Mar 5, 2026

ASRTranscription output contains context.

Status: Open.

#106 In QwenLM/Qwen3-ASR;

· huanchen-1 opened on Mar 4, 2026

Eager nagisa import adds ~3.5s to startup even when ForcedAligner is unused

Status: Open.

#105 In QwenLM/Qwen3-ASR;

· Add1ct1ve opened on Mar 3, 2026

Loading checkpoint failed after finetuning

Status: Open.

#104 In QwenLM/Qwen3-ASR;

· WuZhongdi opened on Mar 2, 2026

Punctuation lost problem for non-English ASR

Status: Open.

#100 In QwenLM/Qwen3-ASR;

· bernardx opened on Feb 25, 2026

Qwen3-ASR Streaming result does not contain punctuations for Arabic

Status: Open.

#94 In QwenLM/Qwen3-ASR;

· XingWang1234 opened on Feb 19, 2026