3 Pull requests opened by 3 people

• SPK recipe for CN-Celeb

  #6126 opened Jun 2, 2025

• Changing order on CI for Whisper installation

  #6129 opened Jun 4, 2025

• Fixed a typo that was causing data leakage.

  #6131 opened Jun 6, 2025

10 Issues closed by 1 person

• Latency of streaming models on CTC only decoding

  #5389 closed Jun 8, 2025

• VITS2

  #5392 closed Jun 8, 2025

• Phoneme recognition recipe

  #5393 closed Jun 8, 2025

• Test LJspeech TTS with random given text.

  #5877 closed Jun 8, 2025

• Computations related to multiply-accumulate (MAC) operations and GFLOPs in speech recognition tasks

  #6041 closed Jun 8, 2025

• Release ESPnet-SpeechLM models on Hugging Face

  #6048 closed Jun 8, 2025

• Issue with MGB-3 FInetuning from MGB2-Recipe

  #6051 closed Jun 8, 2025

• How to inference long audio files in diar_inference

  #5396 closed Jun 5, 2025

• Can I finetune whisper like whisper's training style using prompt as a ASR task

  #5397 closed Jun 5, 2025

• SVS Song Generation Document

  #5402 closed Jun 5, 2025

2 Issues opened by 2 people

• Token list not generating

  #6128 opened Jun 4, 2025

• Авто‑анализ (2025-06-03)

  #6127 opened Jun 3, 2025

34 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Add Squeezeformer based ASR

  #4956 commented on Jun 2, 2025 • 0 new comments

• [Question] How to properly train-infer transformer on padded batches of sequences?

  #4940 commented on Jun 2, 2025 • 0 new comments

• How can I process it faster?

  #4937 commented on Jun 2, 2025 • 0 new comments

• Single channel DCCRN speech enhancement model

  #4930 commented on Jun 2, 2025 • 0 new comments

• New recipe in ESPNet1 with x-vectors and TTS pre-training in Japanese

  #4904 commented on Jun 2, 2025 • 0 new comments

• code for paper "AVOID OVERTHINKING IN SELF-SUPERVISED MODELS FOR SPEECH RECOGNITION"

  #4903 commented on Jun 3, 2025 • 0 new comments

• Voice Conversion(声質変換) recipe

  #4889 commented on Jun 3, 2025 • 0 new comments

• Can we get a cloned voicie in Real Time ?

  #4847 commented on Jun 3, 2025 • 0 new comments

• Question: Voice conversion demo

  #4838 commented on Jun 4, 2025 • 0 new comments

• SeparateSpeech error size mismatch

  #6120 commented on Jun 4, 2025 • 0 new comments

• [OWSM fine-tunening]NameError: name 'tokenize' is not defined

  #6089 commented on Jun 4, 2025 • 0 new comments

• Question: Does ESPnet2 support Internal Language Model Estimation (IMLE) method?

  #4834 commented on Jun 6, 2025 • 0 new comments

• Q: Monitoring loss over each audio

  #4819 commented on Jun 6, 2025 • 0 new comments

• I tryed to train conformer-rnn-transducer on the CSJ dataset.

  #4813 commented on Jun 6, 2025 • 0 new comments

• Question regarding asr_inference_streaming.py

  #4807 commented on Jun 6, 2025 • 0 new comments

• Running the espnet2 example reports an error

  #4810 commented on Jun 6, 2025 • 0 new comments

• Speech2Text error

  #4789 commented on Jun 6, 2025 • 0 new comments

• Extract vocoder from joint model

  #4775 commented on Jun 6, 2025 • 0 new comments

• Is the comment wrong in `label_loader` function?

  #4762 commented on Jun 6, 2025 • 0 new comments

• Request to provide torchscript model

  #4759 commented on Jun 6, 2025 • 0 new comments

• Unable to reproduce egs2/swbd results

  #4602 commented on Jun 6, 2025 • 0 new comments

• How to get correct word from bpe subword

  #4590 commented on Jun 6, 2025 • 0 new comments

• 'freeze-mods' param for finetuning does not handle 'space' well

  #4580 commented on Jun 6, 2025 • 0 new comments

• egs2/aishell2 recipe does not converge

  #4577 commented on Jun 6, 2025 • 0 new comments

• Can't reproduce the result of ESPnet1 in ESPnet2

  #4571 commented on Jun 6, 2025 • 0 new comments

• aishell3 training

  #4561 commented on Jun 6, 2025 • 0 new comments

• NaturalSpeech support

  #4549 commented on Jun 7, 2025 • 0 new comments

• VITS model not uttering some words

  #4546 commented on Jun 7, 2025 • 0 new comments

• RNN-T Decoding - Large Number of Deletions Compared to Transformer/Conformer

  #4540 commented on Jun 7, 2025 • 0 new comments

• Librispeech RNN-T training does not look good with Espnet2

  #4536 commented on Jun 7, 2025 • 0 new comments

• fastspeech2 can inference batch？

  #4532 commented on Jun 7, 2025 • 0 new comments

• RuntimeError: [1] is setting up NCCL communicator and retreiving ncclUniqueId from [0] via c10d key-value store by key '0', but store->get('0') got error: Timeout waiting for key: default_pg/0/0 after 1800000 ms

  #4531 commented on Jun 7, 2025 • 0 new comments

• Decoding process

  #4515 commented on Jun 7, 2025 • 0 new comments

• Training Default Transformer LM fails when using customized corpus

  #4508 commented on Jun 7, 2025 • 0 new comments