shithub: opus

Info  •  Files  •  Log  •  Branches

c7b6935b – Jean-Marc Valin <jmvalin@amazon.com> authored on 2023/05/20 10:21

Add validation for weights blob

0098fe70 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2023/05/19 14:12

Defer calls to run_frame_network() to save CPU

87f9fbc5 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2023/05/18 16:40

Don't preserve owner for model

76c090dc – Jean-Marc Valin <jmvalin@amazon.com> authored on 2023/05/18 13:34

Update model for blob

b9ea8683 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2023/05/17 20:41

Define M_PI when needed

3cd7588d – Jean-Marc Valin <jmvalin@amazon.com> authored on 2023/05/17 20:02

Avoid potential integer wrap-around

580614f0 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2023/05/17 00:09

Make the code C90-compliant

ebbf5721 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2023/05/16 21:08

Use new model struct

71c8a23f – Jean-Marc Valin <jmvalin@amazon.com> authored on 2023/05/16 19:15

Code for building a model struct

cc714cc5 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2023/05/09 23:05

binary weights work in progress

1074e5f0 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2023/05/15 21:07

Properly handle constant tables

2fc6c71d – Jean-Marc Valin <jmvalin@amazon.com> authored on 2023/05/15 11:00

Avoid opus_alloc() conflicting with Opus

0c6c569c – Jean-Marc Valin <jmvalin@amazon.com> authored on 2023/05/14 22:40

Decode RDO-VAE from stack-allocated state

a81a620d – Michael Klingbeil <klingm@amazon.com> authored and Jean-Marc Valin <jmvalin@amazon.com> committed on 2023/05/09 23:09

Remove unused variable that was causing warnings

ea722db2 – Jan Buethe <jbuethe@amazon.de> authored on 2023/02/21 15:22

added -addlpc option to lpcnet_demo

5fbc037f – Jean-Marc Valin <jmvalin@amazon.com> authored on 2023/01/27 13:14

Skip sample rate update on received packets

1bfa0e2b – Jean-Marc Valin <jmvalin@amazon.com> authored on 2023/01/13 10:50

update model

aa474553 – jbuethe <jbuethe@amazon.de> authored on 2023/01/13 06:48

updated torch framework to include quantization

a8673d0e – jbuethe <jbuethe@amazon.de> authored on 2023/01/12 09:15

gru2 -> gruB and dotp included in dump_rdovae

5b9b4381 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/12/20 16:55

Add call to clear FEC buffer

70ce2bd7 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/12/17 21:14

Cleanup, simplifying the no-crossfade PLC case

c0cfcbcf – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/12/19 12:08

update model

e7f23609 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/12/16 07:53

Making it possible to mix DRED with PLC

f178da11 – jbuethe <ubuntu@ip-172-31-44-4.ec2.internal> authored on 2022/12/07 06:29

changed ordering of qframe to reverse in dred_rdovae_decode_qframe

fdb04d0e – jbuethe <jbuethe@amazon.de> authored on 2022/11/23 06:02

added pytorch implementation of RDOVAE

a13aa3a0 – jbuethe <jbuethe@amazon.de> authored on 2022/11/10 10:04

fixed use of quant_levels argument in fec_encoder

861f6739 – jbuethe <jbuethe@amazon.de> authored on 2022/11/09 06:41

added import script for exchange format

ecb5cbcf – jbuethe <jbuethe@amazon.de> authored on 2022/11/08 06:35

added quant-levels argument to dump_rdovae and rdovae_exchange

9859d68b – jbuethe <jbuethe@amazon.de> authored on 2022/11/07 11:14

changed distortion loss weighting back to 0.5, 0.5

0e5a38fa – jbuethe <jbuethe@amazon.de> authored on 2022/11/07 11:13

removed deprecated lambda from fec_encoder

26073861 – jbuethe <jbuethe@amazon.de> authored on 2022/11/04 12:52

fixed scaling/quantization order

a8170986 – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/31 11:21

updated rdovae_exchange

eab9472d – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/31 08:49

added script for exporting RDOVAE weights (external dependency not added yet)

a223122b – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/26 08:42

clean-up

585de8e4 – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/26 06:15

changed data types for r, dead_zone, quant_scale and p0 to opus_uint16

d58faea3 – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/26 05:48

added copyright headers to new files

1f7c39d7 – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/25 09:28

added config include to all new C files

159da408 – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/25 08:59

fixed calculation of p0

818a0496 – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/25 08:15

added re-ordering to dred_rdovae_decode_qframe (frames now in correct order)

23bb11dd – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/25 08:14

added destroy functions for encoder/decoder states

7cac35e6 – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/25 08:13

fixed theta and r values

80383b63 – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/24 07:00

Merge branch 'exp_rdovae6' of https://github.com/xiph/LPCNet into exp_rdovae6

79d4f126 – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/24 07:00

added api functions for retrieving pointers to statistical model parameters

0f044881 – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/24 06:59

fixed type error for statistical model

8accadcc – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/21 15:00

more fixes

d0b1cd88 – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/21 12:35

bugfixes

5b494216 – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/21 11:58

bugfix in dred_rdovae.c

f74bff74 – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/21 11:37

aded api for DRED/RDOVAE to circumvent conflicts between Opus and LPCNet repo

d80f99f7 – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/21 11:33

added void to shut up missing prototype warning

e6390e34 – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/21 08:33

removed compute_dense function (conflict with opus mlp)

1978cc60 – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/21 08:13

refactoring

d1646a68 – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/20 13:27

added NFEC decoder C implementation

ea4d8f54 – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/19 13:18

added statistical model to dump_nfec_model

50966eec – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/19 10:43

bugfixes in nfec encoder

2112f3dd – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/19 06:58

some fixes

c1b357ed – Jan Buethe <jbuethe@amazon.de> authored on 2022/10/18 15:30

first attempt of C implementation of fec encoder (not tested yet due to NEON/DOT_PROD not being separable)

9629ea6a – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/10/13 21:01

Fine-tuning the scripts

0f7fe64d – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/10/04 21:56

Compute FEC features based on loss pattern

89db314e – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/10/04 11:46

Updating fec_encoder.py for recent changes

61459c24 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/10/03 20:27

Change decoder architecture to be like the encoder

79d1a916 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/10/02 08:34

Weighting loss by 1/sqrt(lambda)

0b018637 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/09/30 18:21

Larger range of quantizers

524f8480 – Jan Buethe <jbuethe@amazon.de> authored on 2022/09/30 07:43

removed dump_data delay from total_delay in fec_encoder.py

4c82d3b4 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/09/29 23:53

Completely move quantization out of encoder and decoder

cd0993fd – Jan Buethe <jbuethe@amazon.de> authored on 2022/09/29 17:14

clarifications in help printout

97ffa94d – Jan Buethe <jbuethe@amazon.de> authored on 2022/09/29 17:13

fixed decoder bug (non-quantized input)

589e6741 – jbuethe <jbuethe@amazon.de> authored and Jan Buethe <jbuethe@amazon.de> committed on 2022/09/29 11:53

alignment fix in fec_encoder

a866abe3 – jbuethe <jbuethe@amazon.de> authored and Jan Buethe <jbuethe@amazon.de> committed on 2022/09/28 12:18

ported debug changes from fec_encoder_torch.py

0a2d6dfc – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/09/28 11:34

Use the encoder state as decoder initial state

38dda0f9 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/09/28 11:33

Oops, forgot to run PVQ quantization for the state

b43f077b – jbuethe <jbuethe@amazon.de> authored on 2022/09/27 12:37

corrected offset between decoder output on even/odd latent frames

e4e5958a – jbuethe <jbuethe@amazon.de> authored on 2022/09/27 12:32

Merge branch 'exp_rdovae6' of https://github.com/xiph/LPCNet into exp_rdovae6

01baf1a0 – jbuethe <jbuethe@amazon.de> authored on 2022/09/27 12:31

added missing dead-zone to encode_rdovae.py

be42c3b5 – jbuethe <jbuethe@amazon.de> authored on 2022/09/27 12:29

added fec_encoder.py and corresponding fec_packets.[chpy]

2de335d8 – jbuethe <jbuethe@amazon.de> authored on 2022/09/27 12:22

added fec_encoder.py and fec_packets.[chpy]

ef12c29f – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/09/14 13:04

Update encoder/decoder

405aa7cf – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/09/11 00:13

WIP: training with different alignment

981d06ee – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/09/09 22:18

Refactoring towards multiple offset decoding

a4f7c157 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/09/09 13:03

Stop decimating in the encoder

fdd51eb7 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/09/08 23:51

RepeatVector no longer likes tensor input

8569121f – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/09/07 23:12

RDO-VAE work in progress

b6ac1c78 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/09/23 23:20

FEC hooks in the PLC code

4befd8bb – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/09/23 23:25

Decreasing look-ahead of default model to 1 frame

d45ab6fc – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/09/23 23:19

Move back to tanh for frame rate network

f5c251c5 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/09/19 15:03

Properly align LPC with lookahead in data loader

dd114baf – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/09/15 20:55

Fix causal PLC for models with non-zero lookahead

f3c738d4 – janpbuethe <jbuethe@amazon.de> authored on 2022/09/07 05:03

removed debug prints in dump_lpcnet.py

60450472 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/09/06 20:38

Merge branch 'plc_challenge' into master

340ab308 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/09/06 19:31

model update

920300c5 – janpbuethe <jbuethe@amazon.de> authored and Jean-Marc Valin <jmvalin@amazon.com> committed on 2022/08/31 09:51

Add lpc weighting and model parameter handling

c1da818f – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/05/12 20:43

PLC instructions

ee08ef0a – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/03/17 14:21

Auto-download PLC model weights

27348d1b – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/03/04 11:23

Running new features twice on resync

d2d847bc – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/03/04 11:20

Boosting pitch correlation at inference time

099f0872 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/03/04 11:19

Compensating for DC filter offset during blending

8ca2ccf1 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/02/25 08:57

fix another warning

a814a623 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/02/25 08:56

Silencing a warning

e034b109 – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/02/25 08:55

Biasing for overestimating the pitch correlation

05f02aaa – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/02/24 08:03

Adaptive bias

3982144f – Jean-Marc Valin <jmvalin@amazon.com> authored on 2022/02/23 22:22

Revert PLC state to use Burg in the causal overlap