Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / speechbrain/speechbrain issues and pull requests

#2828 - Create `FetchConfig` for standardizing use of `fetch`

Pull Request - State: open - Opened by pplantinga 5 days ago
Labels: enhancement, refactor

#2827 - Small improvements to core.py for AMP/gradscaler/tqdm

Pull Request - State: open - Opened by Adel-Moumen 5 days ago - 2 comments

#2824 - Changing the default parameters of --compile to make it work more often

Pull Request - State: closed - Opened by TParcollet 6 days ago - 1 comment

#2822 - Fix BEST RQ norm error and efficiency changes

Pull Request - State: closed - Opened by whettenr 9 days ago - 1 comment

#2821 - Bump torchaudio version + add bytes

Pull Request - State: closed - Opened by Adel-Moumen 9 days ago - 2 comments

#2818 - I cannot install speechbrain v1.0.2

Issue - State: open - Opened by Mussaka 12 days ago - 4 comments
Labels: bug

#2817 - [BUG] SpeechBrain version 1.0.2 broke local fetching

Issue - State: open - Opened by daniel-tsiang 14 days ago - 4 comments
Labels: bug

#2815 - InputNormalization with "global" is incorrect

Issue - State: open - Opened by TParcollet 14 days ago
Labels: bug, correctness, confirmed

#2814 - [URGENT] Fix DDP broadcasting

Pull Request - State: closed - Opened by TParcollet 16 days ago - 1 comment
Labels: bug, ready to review, important

#2812 - Quick fix.

Pull Request - State: closed - Opened by TParcollet 18 days ago

#2809 - Inconsistent Device Error in Explicit ASR Inference

Issue - State: open - Opened by cyberso 21 days ago - 3 comments
Labels: bug

#2808 - ASR error rate growth while finetuning

Issue - State: open - Opened by hforghani 22 days ago
Labels: bug

#2807 - Fix #2784 by handling 1-D case for BCE loss

Pull Request - State: closed - Opened by pplantinga 24 days ago - 2 comments
Labels: bug

#2806 - Conformer recipe for LargeScaleASR (code from Samsung AI Center Cambridge)

Pull Request - State: closed - Opened by TParcollet 27 days ago - 1 comment
Labels: ready to review

#2805 - Fix Mimi Bug

Pull Request - State: closed - Opened by poonehmousavi 30 days ago
Labels: bug

#2804 - pre-trained model asr-crdnn-switchboard cannot run inference on cuda

Issue - State: closed - Opened by cyberso about 1 month ago - 2 comments
Labels: bug

#2802 - 25,000 of diverse English ASR data (dataset name hidden) (code from Samsung AI Center Cambridge)

Pull Request - State: open - Opened by TParcollet about 1 month ago - 3 comments
Labels: recipes

#2801 - Convert data download methods to run on main process only

Pull Request - State: closed - Opened by pplantinga about 1 month ago
Labels: refactor

#2800 - Readthedocs now requires a `sphinx.configuration` key

Pull Request - State: closed - Opened by pplantinga about 1 month ago

#2799 - Rotary Position Embedding (RoPE) for ASR (code from Samsung Cambridge)

Pull Request - State: open - Opened by shucongzhang about 1 month ago - 9 comments

#2798 - Test installation fails

Issue - State: closed - Opened by DomDXD about 1 month ago - 3 comments
Labels: bug

#2797 - Use torch.log1p and torch.special.expm1

Pull Request - State: closed - Opened by kit1980 about 1 month ago

#2796 - fix length normalization to prevent ⁇ token generation #2795

Pull Request - State: closed - Opened by gfdb about 1 month ago

#2795 - <eos> token output in `asr-crdnn` models

Issue - State: closed - Opened by pplantinga about 1 month ago
Labels: bug

#2794 - error: subprocess-exited-with-error

Issue - State: closed - Opened by Amogh-GC about 1 month ago - 1 comment
Labels: bug

#2793 - ImportError: There is no such callable as speechbrain.utils.seed_everything: speechbrain 0.5

Issue - State: closed - Opened by saharghannay about 1 month ago - 2 comments
Labels: bug

#2792 - move audio to device when calling load_audio

Pull Request - State: closed - Opened by Adel-Moumen about 2 months ago - 1 comment

#2790 - Streaming recipe for BestRQ

Pull Request - State: open - Opened by Chaanks 2 months ago - 1 comment
Labels: recipes

#2788 - fix unclamped frequencies in drop freq bug

Pull Request - State: closed - Opened by gfdb 2 months ago - 3 comments
Labels: bug

#2787 - unclamped frequency values in dropfreq cause assertion error

Issue - State: closed - Opened by gfdb 2 months ago
Labels: bug

#2786 - `S2SWhisperGreedySearch` not found

Issue - State: closed - Opened by njfamirm 2 months ago - 4 comments
Labels: bug

#2784 - BCE loss fails in 1-D case w/ label smoothing

Issue - State: closed - Opened by pplantinga 2 months ago
Labels: bug

#2783 - [DRAFT] Discrete Speech LMs

Pull Request - State: open - Opened by Adel-Moumen 2 months ago

#2782 - Move files with optional dependencies to integrations folder

Pull Request - State: open - Opened by pplantinga 2 months ago - 11 comments
Labels: correctness

#2781 - Libriheavy (Code from SAIC-Cambridge)

Pull Request - State: closed - Opened by shucongzhang 2 months ago - 2 comments

#2780 - Fix multi-head-attention with adapters

Pull Request - State: closed - Opened by pplantinga 3 months ago - 9 comments
Labels: bug

#2779 - Using Speechbrain MultiheadAttention with PEFT attribute error

Issue - State: closed - Opened by Usanter 3 months ago - 2 comments
Labels: bug

#2775 - Quantization refactor

Pull Request - State: closed - Opened by lucadellalib 3 months ago - 5 comments

#2774 - torchaudio.load is MUCH slower if the backend is ffmpeg.

Issue - State: closed - Opened by TParcollet 3 months ago - 1 comment
Labels: bug

#2774 - torchaudio.load is MUCH slower if the backend is ffmpeg.

Issue - State: open - Opened by TParcollet 3 months ago
Labels: bug

#2773 - Add dropout to ecapa tdnn

Pull Request - State: closed - Opened by pplantinga 3 months ago - 3 comments

#2772 - Alignment with CTC ASR models powered by k2.

Pull Request - State: open - Opened by ZhaoZeyu1995 3 months ago - 6 comments
Labels: enhancement

#2771 - Gammatone dev

Pull Request - State: closed - Opened by naspert 3 months ago

#2768 - increase tolerance for sign flip test #2740

Pull Request - State: closed - Opened by gfdb 3 months ago

#2767 - Add People's Speech (30,000 hours) Conformer ASR (Code from Samsung AI Center Cambridge)

Pull Request - State: closed - Opened by TParcollet 3 months ago - 4 comments
Labels: help wanted, ready to review, recipes

#2764 - SpeechBrain Quantization refactoring

Issue - State: open - Opened by Adel-Moumen 3 months ago - 4 comments
Labels: bug

#2763 - fix safetensors

Pull Request - State: closed - Opened by Adel-Moumen 3 months ago - 4 comments

#2762 - allow local strategy to be passed to collect_files()

Pull Request - State: closed - Opened by mirofedurco 3 months ago - 1 comment
Labels: enhancement

#2760 - Fix Weighted SSL

Pull Request - State: closed - Opened by pplantinga 3 months ago

#2759 - Weighted SSL model not unfreezable

Issue - State: closed - Opened by pplantinga 3 months ago - 1 comment
Labels: bug

#2757 - Italian Hugging Face ASR Models need updates

Issue - State: open - Opened by wredan 3 months ago - 6 comments
Labels: bug

#2756 - SpeechBrain 1.0 Reference needs updated

Issue - State: open - Opened by pplantinga 3 months ago
Labels: documentation, correctness

#2755 - Audio and Music SSL

Pull Request - State: closed - Opened by poonehmousavi 3 months ago - 4 comments
Labels: enhancement

#2754 - Discrete SSL Refactor

Pull Request - State: closed - Opened by Chaanks 3 months ago - 1 comment
Labels: enhancement

#2752 - Add results and links for GigaSpeech RNN-T model

Pull Request - State: closed - Opened by asumagic 3 months ago

#2752 - Add results and links for GigaSpeech RNN-T model

Pull Request - State: closed - Opened by asumagic 3 months ago

#2751 - Add new Audio Tokenziers

Pull Request - State: closed - Opened by poonehmousavi 3 months ago - 16 comments
Labels: enhancement

#2750 - Add Pyre type checker as a pre-commit hook (code from Samsung AI Center Cambridge)

Pull Request - State: closed - Opened by rogiervd 3 months ago - 1 comment

#2750 - Add Pyre type checker as a pre-commit hook (code from Samsung AI Center Cambridge)

Pull Request - State: closed - Opened by rogiervd 3 months ago - 1 comment

#2748 - Feature/update ksponspeech

Pull Request - State: open - Opened by ddwkim 4 months ago

#2748 - Feature/update ksponspeech

Pull Request - State: open - Opened by ddwkim 4 months ago

#2747 - Very WIP triton kernels

Pull Request - State: open - Opened by asumagic 4 months ago

#2746 - only configure the requested logger, not the root logger

Pull Request - State: closed - Opened by benniekiss 4 months ago - 2 comments

#2746 - only configure the requested logger, not the root logger

Pull Request - State: closed - Opened by benniekiss 4 months ago - 2 comments

#2744 - Introducing SpeechBrain Guru on Gurubase.io

Pull Request - State: closed - Opened by kursataktas 4 months ago - 1 comment

#2744 - Introducing SpeechBrain Guru on Gurubase.io

Pull Request - State: closed - Opened by kursataktas 4 months ago - 1 comment

#2740 - `test_sign_flip` does sometimes fail

Issue - State: open - Opened by asumagic 4 months ago - 2 comments
Labels: bug

#2740 - `test_sign_flip` does sometimes fail

Issue - State: closed - Opened by asumagic 4 months ago - 2 comments
Labels: bug

#2738 - Some timers-and-such recipe tests fail

Issue - State: open - Opened by asumagic 4 months ago
Labels: bug

#2736 - Transducer Loss using Triton

Issue - State: open - Opened by Adel-Moumen 4 months ago - 2 comments
Labels: enhancement, performance

#2736 - Transducer Loss using Triton

Issue - State: open - Opened by Adel-Moumen 4 months ago - 2 comments
Labels: enhancement, performance

#2735 - Syntax Bug in Librispeech Whisper Recipe

Issue - State: closed - Opened by matthewkperez 4 months ago - 3 comments
Labels: bug

#2734 - Make calls to "open" deterministic (code from Samsung AI Center Cambridge)

Pull Request - State: closed - Opened by rogiervd 4 months ago - 13 comments
Labels: correctness

#2733 - Investigate custom Triton kernels for depthwise-separable convolutions in Conformer

Issue - State: open - Opened by asumagic 4 months ago - 2 comments
Labels: enhancement, performance

#2730 - Fixes pending for v1.0.2

Pull Request - State: closed - Opened by asumagic 4 months ago

#2729 - Override sub-element of hparams.yaml using command line args?

Issue - State: closed - Opened by Craya 4 months ago - 4 comments
Labels: bug

#2728 - The difference in loss value between "continue" and "interrupt" training

Issue - State: closed - Opened by svecjan 4 months ago - 3 comments
Labels: bug

#2727 - Make `collect_in` optional for `Pretrainer`, disable it by default

Pull Request - State: closed - Opened by asumagic 4 months ago - 2 comments

#2723 - `Pretrained.from_hparams` broke with `savedir is None`

Issue - State: closed - Opened by asumagic 4 months ago
Labels: bug, regression

#2722 - add call on start of fit_batch fn

Pull Request - State: closed - Opened by Adel-Moumen 4 months ago

#2722 - add call on start of fit_batch fn

Pull Request - State: closed - Opened by Adel-Moumen 4 months ago

#2721 - on_fit_batch_start() method not called in the beginning of the fit_batch() method

Issue - State: closed - Opened by tiatariene 4 months ago - 2 comments
Labels: bug

#2721 - on_fit_batch_start() method not called in the beginning of the fit_batch() method

Issue - State: closed - Opened by tiatariene 4 months ago - 2 comments
Labels: bug

#2719 - noisy_wav, lens = batch.noisy_sig?

Issue - State: open - Opened by a897456 4 months ago
Labels: bug

#2719 - noisy_wav, lens = batch.noisy_sig?

Issue - State: open - Opened by a897456 4 months ago - 1 comment
Labels: bug

#2718 - Fixes #2656: Remove EOS from SoundChoice

Pull Request - State: closed - Opened by flexthink 4 months ago - 1 comment