Welcome to mirror list, hosted at ThFree Co, Russian Federation.

gitlab.xiph.org/xiph/opus.git - Unnamed repository; edit this file 'description' to name the repository.
summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2023-09-26changed checkpoint formatexp_pitchdnn1Jan Buethe
2023-09-25Adapting to new data format/modelJean-Marc Valin
2023-09-25Python code for neural pitchKrishna Subramani
2023-09-25Add noise augmentation to pitch dumpingJean-Marc Valin
2023-09-23Add pitch feature computationJean-Marc Valin
2023-09-15Band-limiting the voicing parameterJean-Marc Valin
2023-09-13removed trailing whitespace in farganJan Buethe
Signed-off-by: Jan Buethe <jbuethe@amazon.de>
2023-09-13integrated JM's FFT ada convJan Buethe
Signed-off-by: Jan Buethe <jbuethe@amazon.de>
2023-09-13Remove --has-gain and --passthrough-size argsJean-Marc Valin
Couldn't be used anymore
2023-09-13Add conditioning interpolation, fwconv layerJean-Marc Valin
2023-09-13Only use one frame of pre-loadingJean-Marc Valin
2023-09-13Randomly double the training sequence lengthJean-Marc Valin
Helps with stability with little cost in training speed
2023-09-13Adds skip connectionsJean-Marc Valin
2023-09-135-tap pitch predictorJean-Marc Valin
2023-09-13Also use previous frameJean-Marc Valin
2023-09-13explicit signal gain, explicit pitch predictorJean-Marc Valin
2023-09-13FARGAN initial commit in OpusJean-Marc Valin
Copied/adapted from LPCNet repo
2023-09-12fixJan Buethe
Signed-off-by: Jan Buethe <jbuethe@amazon.de>
2023-09-12added requirements.txt to osceJan Buethe
Signed-off-by: Jan Buethe <jbuethe@amazon.de>
2023-09-12added more enhancement stuffJan Buethe
Signed-off-by: Jan Buethe <jbuethe@amazon.de>
2023-09-05added copyright headersJan Buethe
Signed-off-by: Jan Buethe <jbuethe@amazon.de>
2023-09-05added LPCNet torch implementationJan Buethe
Signed-off-by: Jan Buethe <jbuethe@amazon.de>
2023-09-05brought NoLACE up to dateJan Buethe
Signed-off-by: Jan Buethe <jbuethe@amazon.de>
2023-09-02Fixes vnni macro redefinition with clangMichael Klingbeil
2023-09-01renamed ShapeNet to NoLACEJan Buethe
Signed-off-by: Jan Buethe <jbuethe@amazon.de>
2023-08-10Changing DRED exp. ID so we can reserve 127Jean-Marc Valin
There was a suggestion to use 127 for extending the extension ID space
2023-08-10Don't allocate DRED bits if we don't have enoughJean-Marc Valin
2023-08-10Making it easier to remove DRED experimental IDJean-Marc Valin
When ready, change DRED_EXTENSION_ID to the final ID, remove DRED_EXPERIMENTAL_VERSION completely, and change DRED_EXPERIMENTAL_BYTES to zero (eventually remove it).
2023-08-08improved auto-scaling in wexchangeJan Buethe
2023-08-04Add fwgan_load_model()Jean-Marc Valin
2023-08-04int version for convenienceJean-Marc Valin
2023-08-04Adding some commentsJean-Marc Valin
2023-08-04Continuation now workingJean-Marc Valin
Added required filtering/delay/memory to get continuation
2023-08-03More scaling/filtering to run_fwgan_subframe()Jean-Marc Valin
2023-08-03More FWGAN refactoringJean-Marc Valin
2023-08-03FWGAN refactoringJean-Marc Valin
2023-08-03Applying continuation after first subframeJean-Marc Valin
Continuation matches Python code now
2023-08-02Avoid sin()/cos() calls for pitch referenceJean-Marc Valin
2023-08-02Don't compute linear activation in-placeJean-Marc Valin
saves a few cycles
2023-08-02Add FWGAN to lpcnet_demoexp-fwgan10Jean-Marc Valin
2023-08-02C implementation of FWGANJean-Marc Valin
2023-08-02Skeleton for FWGAN codeJean-Marc Valin
2023-08-02Handle float matrices with multiple of 8 rowsJean-Marc Valin
2023-08-02Add Gated Linear Unit (GLU)Jean-Marc Valin
2023-08-01updated dump_model_weights.pyJan Buethe
2023-08-01added FWGAN weight dumping codeJan Buethe
2023-08-01updadet wexchange version numberJan Buethe
2023-08-01added .copy() to weights in wexchangeJan Buethe
2023-08-01added auto-scaling to wexchangeJan Buethe
2023-07-29fixed wexchange for GRUs without biasJan Buethe
generated by cgit v1.2.3 (git 2.25.1) at 2024-06-03 05:03:41 +0300