Issues

• 有现成的ONNX模型可以使用体验吗？

  wenet-e2e/wekws#219 · Filter by author mallocnew opened on Jun 26, 2026

• 官方文档没有微调、推理、模型导出部分

  wenet-e2e/wekws#214 · Filter by author kunlun122 opened on May 7, 2026
  7

• 流式模型，训练样本长度应该怎么处理

  wenet-e2e/wekws#212 · Filter by author xushangnjlh opened on Feb 9, 2026
  1

• 训练自定义唤醒词，关于唤醒问题

  wenet-e2e/wekws#211 · Filter by author name191 opened on Jan 15, 2026
  2

• DS_TCN Max-Pooling 复现

  wenet-e2e/wekws#210 · Filter by author Nickjiao opened on Dec 30, 2025
  1

• fbank窗函数采用了hamming，和训练python脚本中的kaldi.fbank中的窗函数对不上，导致特征提取对不上

  wenet-e2e/wekws#209 · Filter by author QHenry1990 opened on Nov 12, 2025
  1

• implementation of fsmn

  wenet-e2e/wekws#207 · Filter by author mlxu995 opened on Oct 31, 2025
  1

• 请问工程化runtime的代码，流式推理的fbank是和 torchaudio.kaldi.fbank数值对齐的吗？

  wenet-e2e/wekws#205 · Filter by author shanhaidexiamo opened on Sep 23, 2025
  2

• 关于 ctc 训练的 xiaowen 数据集的 kws 模型

  wenet-e2e/wekws#201 · Filter by author NoirCL opened on Jul 31, 2025
  2

• Making better wakeword datasets

  enhancement
  wenet-e2e/wekws#199 · Filter by author StuartIanNaylor opened on Jul 28, 2025

• 如果我想从头训练一个新的唤醒词，我该如何准备数据呢？

  wenet-e2e/wekws#193 · Filter by author shanhaidexiamo opened on Jun 20, 2025
  4

• 無法成功喚醒

  wenet-e2e/wekws#191 · Filter by author a22802240 opened on Jun 12, 2025
  5

• onnxruntime 在windows上推理

  wenet-e2e/wekws#190 · Filter by author 2217895330 opened on Jun 7, 2025

• fsmn-ctc模型训练后，指令词时间戳不准，超前很多，这个怎么解决？

  wenet-e2e/wekws#188 · Filter by author hzq510473048 opened on May 6, 2025

• 模型训练的时候cv_loss cv_acc =0

  wenet-e2e/wekws#187 · Filter by author 15020096312 opened on Apr 30, 2025
  2

• max_pooling的误唤醒，训练hey siri的模型，单喊一个hey也会唤醒并且得分很高

  wenet-e2e/wekws#186 · Filter by author yuyun2000 opened on Mar 24, 2025
  6

• ctc_fsmn_base 模型的自训练问题---和阿里官方的模型对比

  wenet-e2e/wekws#185 · Filter by author ltcxjtu opened on Feb 10, 2025

• 关于base模型下的学习率设置问题

  wenet-e2e/wekws#184 · Filter by author ltcxjtu opened on Feb 7, 2025

• mtdc speech commands testing

  wenet-e2e/wekws#181 · Filter by author budzianowski opened on Jan 29, 2025

• How to generate different speech rates and pitches through data augmentation on existing data to adapt to generalization

  wenet-e2e/wekws#173 · Filter by author SmallShooter opened on Sep 22, 2024

• torch.distributed.elastic.rendezvous.api.RendezvousConnectionError: The connection to the C10d store has failed. See inner exception for details.

  wenet-e2e/wekws#172 · Filter by author manmushanhe opened on Aug 29, 2024

• About c++11 Using smart Pointers causes source code error protection problems

  wenet-e2e/wekws#170 · Filter by author yanzhe-Xiao opened on Aug 4, 2024
  1

• effect of local shuffle

  wenet-e2e/wekws#167 · Filter by author tomato18463 opened on Jul 5, 2024

• fbank window function discrepancy between training and deployment

  wenet-e2e/wekws#166 · Filter by author tomato18463 opened on Jul 1, 2024
  1

• examples/speechcommand_v1/s0训练的模型在android系统上怎么运行？

  wenet-e2e/wekws#163 · Filter by author taorui-plus opened on May 7, 2024
  2