Commits · 6a81c3a360abbbee44ebe4ff75b5a3cbb821e038 · school / Capstone Design / 01 / AudioLM

Browse files

This project is mirrored from https://github.com/lucidrains/audiolm-pytorch.git. Pull mirroring failed May 17, 2023.
Repository mirroring has been paused due to too many failed attempts. It can be resumed by a project maintainer or owner.
Last successful update May 15, 2023.

Feb 03, 2023
- customize depth of local attention transformer blocks in soundstream · 6a81c3a3
  Phil Wang authored Feb 02, 2023
  
  View commits for tag 0.10.0 0.10.0
  
  6a81c3a3
- 0.9.7 · 513da28c
  Phil Wang authored Feb 02, 2023
  
  View commits for tag 0.9.7 0.9.7
  
  513da28c
- Merge pull request #65 from lucidrains/remove-complex-stft-discriminator · fe382443
  Phil Wang authored Feb 02, 2023
  fe382443
- 0.9.6 · 91d0e1fb
  Phil Wang authored Feb 02, 2023
  
  View commits for tag 0.9.6 0.9.6
  
  91d0e1fb
- Merge pull request #69 from zhvng/forgetful_mask_bug · 42c48b22
  Phil Wang authored Feb 02, 2023
  42c48b22
- bugfix - forgetful mask was not used · e98e856b
  zhvng authored Feb 02, 2023
  
  e98e856b
- helpful error message · 41861d49
  Phil Wang authored Feb 02, 2023
  
  View commits for tag 0.9.5 0.9.5
  
  41861d49
- fix · 2ebe3541
  Phil Wang authored Feb 02, 2023
  
  View commits for tag 0.9.4 0.9.4
  
  2ebe3541
- convenience keyword argument audio_text_condition, just have researchers set... · 49cc3923
  Phil Wang authored Feb 02, 2023
  View commits for tag 0.9.3 0.9.3
  
  49cc3923
- make sure one can detect easily whether audio conditioner is present for AudioLM · b3a763c2
  Phil Wang authored Feb 02, 2023
  
  View commits for tag 0.9.2 0.9.2
  
  b3a763c2
- each hierarchical transformer will have its own set of conditioning embeddings... · 15151252
  Phil Wang authored Feb 02, 2023
  View commits for tag 0.9.1 0.9.1
  
  15151252
- add ability to condition on joint text-audio embeddings from mulan, over at musiclm-pytorch · 2e654ba8
  Phil Wang authored Feb 02, 2023
  
  View commits for tag 0.9.0 0.9.0
  
  2e654ba8
- use a base class to enforce what the trainer can accept, for musiclm · 04bed934
  Phil Wang authored Feb 02, 2023
  
  View commits for tag 0.8.3 0.8.3
  
  04bed934
- loosen restrictions around what is passed in for transformers into the... · 88bf4dfb
  Phil Wang authored Feb 02, 2023
  88bf4dfb
- prepare for musiclm inference and also address... · 8db0afee
  Phil Wang authored Feb 02, 2023
  View commits for tag 0.8.1 0.8.1
  
  8db0afee
Feb 02, 2023
- 0.7.9 · 6919830b
  Phil Wang authored Feb 01, 2023
  
  View commit 6919830b 2 tags
  
  6919830b
- Merge pull request #68 from djqualia/patch-10 · 441aa72b
  Phil Wang authored Feb 01, 2023
  441aa72b
- Convert stereo/multi-channel audio to mono · ab4fc99a
  djqualia authored Feb 01, 2023
  ab4fc99a
- less is more · 9b62899e
  Phil Wang authored Feb 01, 2023
  
  9b62899e
- appreciation · ba7dcd68
  Phil Wang authored Feb 01, 2023
  
  ba7dcd68
- default back to distributed friendly soundstream, lower default learning rate... · 55da605a
  Phil Wang authored Feb 01, 2023
  View commits for tag 0.7.8 0.7.8
  
  55da605a
Feb 01, 2023
- potentially fix https://github.com/lucidrains/audiolm-pytorch/issues/64 · 47ee1c39
  Phil Wang authored Jan 31, 2023
  
  View commits for tag 0.7.7 0.7.7
  
  47ee1c39
- get rid of a warning message · edb8f779
  Phil Wang authored Jan 31, 2023
  
  edb8f779
- default back to stft discriminator using torch.complex64, due to... · fe97568c
  Phil Wang authored Jan 31, 2023
  View commits for tag 0.7.6 0.7.6
  
  fe97568c
Jan 31, 2023
- Merge pull request #63 from LWprogramming/demo_ipynb · c6c8299a
  Phil Wang authored Jan 30, 2023
  c6c8299a
- Add demo Jupyter notebook to run e2e · e65a4333
  Leon Wu authored Jan 30, 2023
  
  e65a4333
Jan 29, 2023
- appreciation · a5b4b990
  Phil Wang authored Jan 28, 2023
  
  a5b4b990
- just set as default and wait for a researcher to raise an issue · d7623ab3
  Phil Wang authored Jan 28, 2023
  
  View commits for tag 0.7.5 0.7.5
  
  d7623ab3
Jan 28, 2023
- attempt to fix distributed training for soundstream in... · dfe9a1e8
  Phil Wang authored Jan 27, 2023
  View commits for tag 0.7.4 0.7.4
  
  dfe9a1e8
- update · 49499085
  Phil Wang authored Jan 27, 2023
  
  49499085
Jan 26, 2023
- bring back the complex stft discriminator, to try to figure out where the... · 2ebfd1c1
  Phil Wang authored Jan 25, 2023
  View commits for tag 0.7.3 0.7.3
  
  2ebfd1c1
Jan 25, 2023
- revert changes to elu activation placement, as researcher @yigityu is hitting convergence issues · e6eaabdb
  Phil Wang authored Jan 24, 2023
  
  View commits for tag 0.7.2 0.7.2
  
  e6eaabdb
Jan 21, 2023
- readme · d06fcf7e
  Phil Wang authored Jan 20, 2023
  
  d06fcf7e
- add learned multi-headed exponential moving average as an option for soundstream · ed10de42
  Phil Wang authored Jan 20, 2023
  
  View commits for tag 0.7.1 0.7.1
  
  ed10de42
- fix when turning off local attention for sound stream · 455def25
  Phil Wang authored Jan 20, 2023
  
  View commits for tag 0.6.3 0.6.3
  
  455def25
- switch to continuous positional bias, for the length extrapolation at inference time · d5890a24
  Phil Wang authored Jan 20, 2023
  
  View commits for tag 0.6.2 0.6.2
  
  d5890a24
- make dilation order in encoder and decoder configurable in soundstream · 8e3d1979
  Phil Wang authored Jan 20, 2023
  
  View commits for tag 0.6.1 0.6.1
  
  8e3d1979
Jan 20, 2023
- use the same activation placements as encodec, suggested by @inspirit · 36c39540
  Phil Wang authored Jan 20, 2023
  
  View commits for tag 0.6.0 0.6.0
  
  36c39540
Jan 19, 2023
- make sure to address the discriminator gradient issues uncovered by... · daeedb27
  Phil Wang authored Jan 18, 2023
  View commits for tag 0.5.1 0.5.1
  
  daeedb27
- make sure local attention blocks are followed by feedforward · ec642cd9
  Phil Wang authored Jan 18, 2023
  
  View commits for tag 0.5.0 0.5.0
  
  ec642cd9