Commits · 5a29ca93acc1b165498e1d8938d7bdfdf89c167a · Xiph.Org / Opus

Oct 30, 2023
- Don't use reserved identifiers for include guards · 61fb3b16
  Jean-Marc Valin authored 1 year ago
  
  61fb3b16
Jun 22, 2023
- Remove trailing whitespace in dnn · f36685fc
  Marcus Asteborg authored 1 year ago
  
  f36685fc
May 22, 2023
- Convert PLC weights to blob format · 98da3350
  Jean-Marc Valin authored 1 year ago
  
  98da3350
May 20, 2023
- Add validation for weights blob · c7b6935b
  Jean-Marc Valin authored 1 year ago
  
  c7b6935b
May 17, 2023
- Use new model struct · ebbf5721
  Jean-Marc Valin authored 1 year ago
  
  ebbf5721
- Code for building a model struct · 71c8a23f
  Jean-Marc Valin authored 1 year ago
  
  71c8a23f
May 16, 2023
- binary weights work in progress · cc714cc5
  Jean-Marc Valin authored 1 year ago
  
  cc714cc5
Jan 12, 2023
- gru2 -> gruB and dotp included in dump_rdovae · a8673d0e
  Jan Buethe authored 2 years ago
  
  a8673d0e
Dec 19, 2022
- update model · c0cfcbcf
  Jean-Marc Valin authored 2 years ago
  
  New model that can handle absent Burg features
  c0cfcbcf
Nov 10, 2022
- fixed use of quant_levels argument in fec_encoder · a13aa3a0
  Jan Buethe authored 2 years ago
  
  a13aa3a0
Nov 09, 2022
- added import script for exchange format · 861f6739
  Jan Buethe authored 2 years ago
  
  861f6739
Nov 08, 2022
- added quant-levels argument to dump_rdovae and rdovae_exchange · ecb5cbcf
  Jan Buethe authored 2 years ago
  
  ecb5cbcf
Nov 07, 2022
- changed distortion loss weighting back to 0.5, 0.5 · 9859d68b
  Jan Buethe authored 2 years ago
  
  9859d68b
- removed deprecated lambda from fec_encoder · 0e5a38fa
  Jan Buethe authored 2 years ago
  
  0e5a38fa
Nov 04, 2022
- fixed scaling/quantization order · 26073861
  Jan Buethe authored 2 years ago
  
  26073861
Oct 31, 2022
- updated rdovae_exchange · a8170986
  Jan Buethe authored 2 years ago
  
  a8170986
- added script for exporting RDOVAE weights (external dependency not added yet) · eab9472d
  Jan Buethe authored 2 years ago
  
  eab9472d
Oct 26, 2022
- changed data types for r, dead_zone, quant_scale and p0 to opus_uint16 · 585de8e4
  Jan Buethe authored 2 years ago
  
  585de8e4
- added copyright headers to new files · d58faea3
  Jan Buethe authored 2 years ago
  
  d58faea3
Oct 25, 2022
- fixed calculation of p0 · 159da408
  Jan Buethe authored 2 years ago
  
  159da408
- fixed theta and r values · 7cac35e6
  Jan Buethe authored 2 years ago
  
  7cac35e6
Oct 24, 2022
- fixed type error for statistical model · 0f044881
  Jan Buethe authored 2 years ago
  
  0f044881
Oct 21, 2022
- more fixes · 8accadcc
  Jan Buethe authored 2 years ago
  
  8accadcc
- refactoring · 1978cc60
  Jan Buethe authored 2 years ago
  
  1978cc60
Oct 20, 2022
- added NFEC decoder C implementation · d1646a68
  Jan Buethe authored 2 years ago
  
  d1646a68
Oct 19, 2022
- added statistical model to dump_nfec_model · ea4d8f54
  Jan Buethe authored 2 years ago
  
  ea4d8f54
Oct 18, 2022
- first attempt of C implementation of fec encoder (not tested yet due to... · c1b357ed
  Jan Buethe authored 2 years ago
  
  first attempt of C implementation of fec encoder (not tested yet due to NEON/DOT_PROD not being separable)
  c1b357ed
Oct 14, 2022
- Fine-tuning the scripts · 9629ea6a
  Jean-Marc Valin authored 2 years ago
  
  9629ea6a
Oct 05, 2022
- Compute FEC features based on loss pattern · 0f7fe64d
  Jean-Marc Valin authored 2 years ago
  
  0f7fe64d
Oct 04, 2022
- Updating fec_encoder.py for recent changes · 89db314e
  Jean-Marc Valin authored 2 years ago
  
  89db314e
- Change decoder architecture to be like the encoder · 61459c24
  Jean-Marc Valin authored 2 years ago
  
  61459c24
Oct 02, 2022
- Weighting loss by 1/sqrt(lambda) · 79d1a916
  Jean-Marc Valin authored 2 years ago
  
  79d1a916
Oct 01, 2022
- Larger range of quantizers · 0b018637
  Jean-Marc Valin authored 2 years ago
  
  0b018637
Sep 30, 2022
- removed dump_data delay from total_delay in fec_encoder.py · 524f8480
  Jan Buethe authored 2 years ago
  
  524f8480
- Completely move quantization out of encoder and decoder · 4c82d3b4
  Jean-Marc Valin authored 2 years ago
  
  4c82d3b4
Sep 29, 2022
- clarifications in help printout · cd0993fd
  Jan Buethe authored 2 years ago
  
  cd0993fd
- fixed decoder bug (non-quantized input) · 97ffa94d
  Jan Buethe authored 2 years ago
  
  97ffa94d
- alignment fix in fec_encoder · 589e6741
  Jan Buethe authored 2 years ago and Jan Buethe committed 2 years ago
  
  589e6741
- ported debug changes from fec_encoder_torch.py · a866abe3
  Jan Buethe authored 2 years ago and Jan Buethe committed 2 years ago
  
  a866abe3
Sep 28, 2022
- Use the encoder state as decoder initial state · 0a2d6dfc
  Jean-Marc Valin authored 2 years ago
  
  Helps reduce the error on the most recent frames
  0a2d6dfc