GitHub / speechbrain/speechbrain / commits
A PyTorch-based Speech Toolkit
| SHA | Message | Author | Date | Stats |
|---|---|---|---|---|
| e5cb1f65 |
Fix #3033: Fix closure variable bug in classification_error (#3061)
Signed-off-by: nandanadileep <n****9@g****m> |
Nandana Dileep <1****p@u****m>
Committed by: GitHub <n****y@g****m> |
2 months ago | |
| d0d15bbb |
Fix incorrect documented defaults for HF wrapper freeze/output_norm (#3060) (...
Co-authored-by: Deeksha Manjunath <1****A@u****m> |
Deeksha Manjunath <d****h@n****m>
Committed by: GitHub <n****y@g****m> |
2 months ago | |
| 2827a48d |
docs: fix incorrect parameter description in load_data_json (#3053)
Co-authored-by: Parcollet Titouan <p****n@g****m> |
Bimochan raj kunwar <1****n@u****m>
Committed by: GitHub <n****y@g****m> |
3 months ago | |
| 8a89ebad |
Fix device handling for MPS (#3055)
Co-authored-by: flexthink <f****k@u****m> |
flexthink <1****k@u****m>
Committed by: GitHub <n****y@g****m> |
4 months ago | |
| da704dd9 | FocalCodec-Stream [ICASSP 2026] (#3050) |
Luca Della Libera <3****b@u****m>
Committed by: GitHub <n****y@g****m> |
4 months ago | |
| 9c826ef3 | Add hooks to ensure saving of kmeans model works correctly (#3047) |
Peter Plantinga <p****r@p****m>
Committed by: GitHub <n****y@g****m> |
5 months ago | |
| 638ffd6c | Bump version from 1.0.4 to 1.1.0 |
Adel Moumen <8****n@u****m>
Committed by: GitHub <n****y@g****m> |
5 months ago | |
| 72534a74 | merge Merge branch 'main' into develop |
Adel Moumen <8****n@u****m>
Committed by: GitHub <n****y@g****m> |
5 months ago | |
| 599911f5 | Bump version from 1.0.3 to 1.0.4 |
Adel Moumen <8****n@u****m>
Committed by: GitHub <n****y@g****m> |
5 months ago | |
| 8b133b65 | Recipe tests new release (#3042) |
Adel Moumen <8****n@u****m>
Committed by: GitHub <n****y@g****m> |
5 months ago | |
| aca5e41a |
Adding SENSE models (#2998)
Co-authored-by: Adel Moumen <8****n@u****m>, Copilot <1****t@u****m> |
bouziane maryem <1****e@u****m>
Committed by: GitHub <n****y@g****m> |
6 months ago | |
| 3eb32f63 | Fix author name typo: Abous-Rjeili → Abou-Rjeili (#3038) |
Georges <5****r@u****m>
Committed by: GitHub <n****y@g****m> |
6 months ago | |
| 47d71953 | Fix mutable default arguments in function signatures (#3034) |
Harikrishna KP <h****2@g****m>
Committed by: GitHub <n****y@g****m> |
7 months ago | |
| d48d36cb | support torch>=2.9.0 (#3032) |
Emmanuel Schmidbauer <e****r@g****m>
Committed by: GitHub <n****y@g****m> |
7 months ago | |
| a601051f | Add missing mean_stat_per_model method to StatObject_SB (#3029) |
Harikrishna KP <h****2@g****m>
Committed by: GitHub <n****y@g****m> |
7 months ago | |
| 1f98c521 |
Remove CTC CUDA + Move transducer loss in integrations (#3028)
Co-authored-by: Copilot <1****t@u****m> |
Adel Moumen <8****n@u****m>
Committed by: GitHub <n****y@g****m> |
7 months ago | |
| 39ef3587 | SpeechLLM LibriSpeech recipe (#2885) |
Adel Moumen <8****n@u****m>
Committed by: GitHub <n****y@g****m> |
7 months ago | |
| de1c94d8 |
Update scorer.py (#2617)
Co-authored-by: Peter Plantinga <p****r@p****e> |
Hemant Yadav <r****h@g****m>
Committed by: GitHub <n****y@g****m> |
7 months ago | |
| 3828d3b8 |
Fix typo in literature reference (#3022)
Co-authored-by: Adel Moumen <8****n@u****m> |
Ibrahim Dahiru Dahiru <3****r@u****m>
Committed by: GitHub <n****y@g****m> |
7 months ago | |
| c3c34d3b |
add tests + doc + fix recipes + new get_available_cpu_count fn (#3025)
Co-authored-by: Copilot <1****t@u****m> |
Adel Moumen <8****n@u****m>
Committed by: GitHub <n****y@g****m> |
7 months ago | |
| 0202ec53 | Update CI requirements (#3023) |
Adel Moumen <8****n@u****m>
Committed by: GitHub <n****y@g****m> |
7 months ago | |
| f97c1b2f | Fix text_file assignment - SentencePiece (#3016) |
Peter Plantinga <p****r@p****m>
Committed by: GitHub <n****y@g****m> |
7 months ago | |
| ba4f674b |
Fix duplicate mic position in beamforming tutorial (#3019)
Co-authored-by: Claude Opus 4.5 <n****y@a****m> |
Adel Moumen <8****n@u****m>
Committed by: GitHub <n****y@g****m> |
7 months ago | |
| 264b642d |
Replace deprecated torch.cuda.amp.custom_fwd with torch.amp.custom_fw (#2922)
Co-authored-by: Adel Moumen <8****n@u****m>, Peter Plantinga <p****r@p****m>, copilot-swe-agent[bot] <1****t@u****m> |
Omid Saghatchian <o****n@g****m>
Committed by: GitHub <n****y@g****m> |
9 months ago | |
| 1e32ef89 | ksponspeech remove (#3013) |
Adel Moumen <8****n@u****m>
Committed by: GitHub <n****y@g****m> |
9 months ago | |
| c72fff46 | Add init files so pages get picked up for docs (#3011) |
Peter Plantinga <p****r@p****m>
Committed by: GitHub <n****y@g****m> |
9 months ago | |
| c94d5e29 |
Implement per-key padding configuration in PaddedBatch (#3008)
Co-authored-by: Copilot <1****t@u****m> |
Adel Moumen <8****n@u****m>
Committed by: GitHub <n****y@g****m> |
9 months ago | |
| fac03671 | Fix local_rank when using single process (#3006) |
Adel Moumen <8****n@u****m>
Committed by: GitHub <n****y@g****m> |
9 months ago | |
| b3e80b48 | HDF5 integration not loading/caching etc. (#3007) |
Adel Moumen <8****n@u****m>
Committed by: GitHub <n****y@g****m> |
9 months ago | |
| 68996739 |
Caching: add compression + filename + closing/loading (#3005)
Co-authored-by: Copilot <1****t@u****m> |
Adel Moumen <8****n@u****m>
Committed by: GitHub <n****y@g****m> |
9 months ago | |
| 637f0a55 |
FocalCodec [NeurIPS 2025] (#3000)
Co-authored-by: Mirco Ravanelli <m****i@g****m> |
Luca Della Libera <3****b@u****m>
Committed by: GitHub <n****y@g****m> |
9 months ago | |
| db2db458 |
Rename FetchConfig.use_auth_token to token and pass token to HF hub (#2996)
Co-authored-by: Peter Plantinga <p****r@p****m> |
bnmy6581 <4****n@u****m>
Committed by: GitHub <n****y@g****m> |
9 months ago | |
| c3b5fcea |
fix issue with --arg value and --arg=value (#2999)
Co-authored-by: Copilot <1****t@u****m> |
Adel Moumen <8****n@u****m>
Committed by: GitHub <n****y@g****m> |
9 months ago | |
| 18b41a45 | Update versions of python due to 3.9 EOL (#3001) |
Peter Plantinga <p****r@p****m>
Committed by: GitHub <n****y@g****m> |
9 months ago | |
| 2b3f4f44 |
SGMSE Voicebank Speech Enhancement Recipe (#2947)
Co-authored-by: Jonas Rochdi <1****i@i****e>, Peter Plantinga <p****r@p****m> |
Jonas <1****i@u****m>
Committed by: GitHub <n****y@g****m> |
9 months ago | |
| 2427785d |
Replace torchaudio I/O with soundfile-based audio_io wrapper (#2989)
Co-authored-by: copilot-swe-agent[bot] <1****t@u****m>, pplantinga <4****a@u****m>, Peter Plantinga <p****r@p****e>, Parcollet Titouan <p****n@g****m>, Peter Plantinga <p****r@p****m> |
Copilot <1****t@u****m>
Committed by: GitHub <n****y@g****m> |
9 months ago | |
| a125f316 | Add functions for running code once per node (#2992) |
Peter Plantinga <p****r@p****m>
Committed by: GitHub <n****y@g****m> |
9 months ago | |
| b7afd035 | fix speed perturb inversion (#2994) |
gianfranco <6****b@u****m>
Committed by: GitHub <n****y@g****m> |
10 months ago | |
| c6d2a8b1 |
Fix for Tracing ECAPA-TDNN Embedding Model (#2982)
Co-authored-by: Parcollet Titouan <p****n@g****m> |
Jordan Zangara <j****g@p****m>
Committed by: GitHub <n****y@g****m> |
10 months ago | |
| 43c8ad1b |
Adapting transducer greedy decoding (#2975)
Co-authored-by: Parcollet Titouan <p****n@g****m>, Youness Dkhissi <y****i@o****m> |
Youness Dkhissi <8****i@u****m>
Committed by: GitHub <n****y@g****m> |
10 months ago | |
| dadc1d3b | Feature caching proposal: CachedDynamicItem (#2985) |
Peter Plantinga <p****r@p****m>
Committed by: GitHub <n****y@g****m> |
10 months ago | |
| 927530fa |
Bandaid fix for torchaudio 2.9+ compatibility (#2988)
Co-authored-by: Peter Plantinga <p****r@p****e> |
Oscar Friedman <o****r@f****m>
Committed by: GitHub <n****y@g****m> |
10 months ago | |
| 279a09a1 |
Fix #2973 bug with period in path (#2974)
Co-authored-by: Parcollet Titouan <p****n@g****m> |
Peter Plantinga <p****r@p****m>
Committed by: GitHub <n****y@g****m> |
11 months ago | |
| ac674c0d |
Bump torch from 2.7.1 to 2.8.0 in /docs (#2976)
Co-authored-by: dependabot[bot] <4****]@u****m> Signed-off-by: dependabot[bot] <s****t@g****m> |
dependabot[bot] <4****]@u****m>
Committed by: GitHub <n****y@g****m> |
11 months ago | |
| f3b75208 | Update version constraints for torch and torchaudio (#2978) |
Adel Moumen <8****n@u****m>
Committed by: GitHub <n****y@g****m> |
11 months ago | |
| 0dbaf795 |
Adopt pyproject and move to Ruff (#2946)
Co-authored-by: Cheeky Sparrow <c****k@c****m> |
Parcollet Titouan <p****n@g****m>
Committed by: GitHub <n****y@g****m> |
11 months ago | |
| 7d214a38 |
fix unnecessary pos_emb for RoPE (#2963)
Co-authored-by: Shucong Zhang/Embedded AI /SRUK/Engineer/Samsung Electronics <s****g@s****t> |
shucongzhang <1****g@u****m>
Committed by: GitHub <n****y@g****m> |
about 1 year ago | |
| 308527c0 | Use epoch loop loader for transfer to correctly handle end_of_epoch (#2965) |
Peter Plantinga <p****r@p****m>
Committed by: GitHub <n****y@g****m> |
about 1 year ago | |
| 96b00ca4 |
SpeechLLM (with LLaMA) and Conformer recipe for speech translation on CoVoST ...
Co-authored-by: Adel Moumen <a****o@g****m>, Adel Moumen <8****n@u****m> |
Parcollet Titouan <p****n@g****m>
Committed by: GitHub <n****y@g****m> |
about 1 year ago | |
| 33474cb3 |
Add recipe installation instructions to installation.md (#2948)
Co-authored-by: Adel Moumen <8****n@u****m> |
Peter Plantinga <p****r@p****m>
Committed by: GitHub <n****y@g****m> |
about 1 year ago | |
| 1fedcefd | Fix missing defaults in run options (#2952) |
Peter Plantinga <p****r@p****m>
Committed by: GitHub <n****y@g****m> |
about 1 year ago | |
| ec142536 | Fix forced alignment tutorial link to google colab (#2951) |
Peter Plantinga <p****r@p****m>
Committed by: GitHub <n****y@g****m> |
about 1 year ago | |
| 13cd75ca |
Alignment with CTC ASR models powered by k2. (#2772)
Co-authored-by: Peter Plantinga <p****r@p****e> |
Zeyu Zhao <z****5@g****m>
Committed by: GitHub <n****y@g****m> |
about 1 year ago | |
| 246a952c |
Saveable Generator: initial import (#2937)
Co-authored-by: flexthink <f****k@u****m> |
flexthink <1****k@u****m>
Committed by: GitHub <n****y@g****m> |
about 1 year ago | |
| 2eb47df9 |
Make run_on_main and main_process_only return the result to all proce… (#2943)
Co-authored-by: Rogier van Dalen <r****n@s****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
about 1 year ago | |
| 900f02ca |
Solves #2846 Refactor run_opts into a @dataclass (#2855)
Co-authored-by: Peter Plantinga <p****r@p****e> |
Nouran Ali <n****d@g****m>
Committed by: GitHub <n****y@g****m> |
about 1 year ago | |
| bca9c766 |
Bump torch from 2.6.0 to 2.7.1 in /docs (#2939)
Co-authored-by: dependabot[bot] <4****]@u****m> Signed-off-by: dependabot[bot] <s****t@g****m> |
dependabot[bot] <4****]@u****m>
Committed by: GitHub <n****y@g****m> |
about 1 year ago | |
| c75ab548 | Fix issue 'GlobalNorm with DDP' (#2934) |
Jan Svec <3****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 1 year ago | |
| 4f4b49c1 | Create `FetchConfig` for standardizing use of `fetch` (#2828) |
Peter Plantinga <p****r@p****m>
Committed by: GitHub <n****y@g****m> |
about 1 year ago | |
| 69161753 | Re-ran tutorial to remove errors (#2925) |
Peter Plantinga <p****r@p****m>
Committed by: GitHub <n****y@g****m> |
about 1 year ago | |
| 075d5f24 |
Librilight data preparation for SpeechBrain SSL (code from Samsung AI Center ...
Co-authored-by: Shucong Zhang/Embedded AI /SRUK/Engineer/Samsung Electronics <s****g@s****t>, Parcollet Titouan <p****n@g****m> |
shucongzhang <1****g@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 20ee80a6 |
Streaming recipe for BestRQ (#2790)
Co-authored-by: Adel Moumen <a****o@g****m>, Titouan Parcollet <p****n@g****m> |
Jarod <j****5@g****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 068a41bb |
Move files with optional dependencies to integrations folder (#2782)
Co-authored-by: Adel Moumen <a****o@g****m> |
Peter Plantinga <p****r@p****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| e2696d98 | Renaming largescale ASR to the Loquacious Set (#2928) |
Parcollet Titouan <p****n@g****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| f703c6ea | docs: Fix contributing.md (#2923) |
Omid Saghatchian <o****n@g****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| d10acec9 | Fix Black error introduced in #2871 | Peter Plantinga <p****r@p****e> | over 1 year ago | |
| fe9178c6 |
fix bug in reverberate with rescale_amp (#2871)
Co-authored-by: Parcollet Titouan <p****n@g****m> |
ofiryaish <o****h@g****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 0ff5c0c8 |
Change test for make_padding_mask to expose mistake (#2895)
Co-authored-by: Rogier van Dalen <r****n@s****m>, Peter Plantinga <p****r@p****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| ba9364af |
Add test for various length mask functions (code from Samsung, AI Center, Cam...
Co-authored-by: Rogier van Dalen <r****n@s****m>, Adel Moumen <8****n@u****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| f7699261 |
Fix type annotations (from Samsung, AI Center, Cambridge) (#2911)
Co-authored-by: Rogier van Dalen <r****n@s****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 1d5ed136 |
Change location from which to import tqdm (#2921)
Co-authored-by: Rogier van Dalen <r****n@s****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| c7ba0feb |
Use "cls" instead of "self" in classmethod (#2908)
Co-authored-by: Rogier van Dalen <r****n@s****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| aebe799b |
Remove deprecated abstractproperty (from Samsung, AI Center, Cambridge) (#2915)
Co-authored-by: Rogier van Dalen <r****n@s****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 6c4e6fc9 |
Make Scores.__repr__ actually return something (#2897)
Co-authored-by: Rogier van Dalen <r****n@s****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 3f2a3dbd |
Remove unused value (from Samsung, AI Center, Cambridge) (#2914)
Co-authored-by: Rogier van Dalen <r****n@s****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 1f6192fa |
Replace axis= by dim= for calls to Torch (from Samsung, AI Center, Cambridge)...
Co-authored-by: Rogier van Dalen <r****n@s****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| e17841af |
Do not forget to raise ValueError (from Samsung, AI Center, Cambridge) (#2909)
Co-authored-by: Rogier van Dalen <r****n@s****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 7d897730 |
Migrate to modern Python Logger API (#2900)
Signed-off-by: Emmanuel Ferdman <e****n@g****m> |
Emmanuel Ferdman <e****n@g****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| edc8db40 |
Account for all combinations of arguments in IterativeCSVWriter.write (from S...
Co-authored-by: Rogier van Dalen <r****n@s****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 4e47d940 |
Make f_max integer (from Samsung, AI Center, Cambridge) (#2918)
Co-authored-by: Rogier van Dalen <r****n@s****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 4ade5319 |
Make arpa_to_fst work when arguments are str (from Samsung, AI Center, Cambri...
Co-authored-by: Rogier van Dalen <r****n@s****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 5d81ad12 |
Update handling of dtype in read_pkl to torch.Tensor (from Samsung, AI Center...
Co-authored-by: Rogier van Dalen <r****n@s****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 923a5686 |
Use correct signature for torch.tensor.type() (from Samsung, AI Center, Cambr...
Co-authored-by: Rogier van Dalen <r****n@s****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 453639f5 |
Make types correct (from Samsung, AI Center, Cambridge) (#2912)
Co-authored-by: Rogier van Dalen <r****n@s****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 20b18a01 |
Fix bug constructing ValueError in Conv2d (from Samsung, AI Center, Cambridge...
Co-authored-by: Rogier van Dalen <r****n@s****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 4ebffc6a |
Fix bug in ModuleList.insert (from Samsung, AI Center, Cambridge) (#2903)
Co-authored-by: Rogier van Dalen <r****n@s****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 2d014fe3 |
Ensure that x is not unbound in WordEmbeddingEncoder (from Samsung, AI Center...
Co-authored-by: Rogier van Dalen <r****n@s****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 70dc8f77 |
Fix argument to rec_loss in (Variational)AutoencoderLoss (from Samsung AI Cen...
Co-authored-by: Rogier van Dalen <r****n@s****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| e246ea66 |
Bump version of pyroomacoustics for recipes/DNS/enhancement (#2899)
Co-authored-by: Rogier van Dalen <r****n@s****m> |
Rogier van Dalen <r****d@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 9e068c4e |
Fix input normalization and global normalization variance calculation (#2835)
Co-authored-by: Rogier van Dalen <r****n@s****m>, Parcollet Titouan <p****n@g****m> |
Peter Plantinga <p****r@p****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 256d25f7 | Avoid running tests on draft PRs (#2890) |
Peter Plantinga <p****r@p****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| eb0532a2 |
Fix dropout at inference (#2889)
Co-authored-by: Peter Plantinga <p****r@p****m> |
Adel Moumen <8****n@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 192aa38c | Attempt fix for input_norm test (#2887) |
Peter Plantinga <p****r@p****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 806a7ab1 | fix convolutions docstrings (#2883) |
gianfranco <6****b@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| e2c4ba47 | Attempt to fix `self.device` for AMP (#2882) |
Adel Moumen <8****n@u****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 00b981c7 | Reforging LLaMA lobe (code from Samsung AI Center-Cambridge) (#2850) |
Parcollet Titouan <p****n@g****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| b9176fb2 | Fix broken links due to change in voxlingua107 location (#2877) |
Peter Plantinga <p****r@p****m>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| 31c1e329 | merge Merge branch 'develop' | Peter Plantinga <p****r@p****m> | over 1 year ago | |
| b2d24131 | Update version | Peter Plantinga <p****r@p****m> | over 1 year ago | |
| c5424356 | Update version | Peter Plantinga <p****r@p****m> | over 1 year ago |