Age | Commit message (Collapse) | Author | |
---|---|---|---|
2023-11-16 | Adding RTCD for compute_activation()exp_dnn_rtcd4exp_dnn_rtcd2 | Jean-Marc Valin | |
2023-11-16 | Adding RTCD for DNN code | Jean-Marc Valin | |
Starting with compute_linear() | |||
2023-11-15 | Using sparse GRUs in DRED decoder | Jean-Marc Valin | |
Saves ~270 kB of weights in the decoder | |||
2023-11-11 | Fix non-AVX builds | Jean-Marc Valin | |
2023-11-09 | Dumping RDOVAE stats from XML | Jean-Marc Valin | |
2023-11-08 | DRED: quantize scale and dead zone to 8 bits | Jean-Marc Valin | |
2023-11-07 | added weight export script for LACE/NoLACE | Jan Buethe | |
2023-11-07 | added transposed 1d convolutions to wexchange | Jan Buethe | |
2023-11-07 | DRED code cleanup | Jean-Marc Valin | |
Removing some indirections | |||
2023-11-07 | Split stats in two and remove useless dimensions | Jean-Marc Valin | |
2023-11-06 | updated moc to match results in ietf118 presentation | Jan Buethe | |
2023-11-06 | DRED: quantize r and p0 parameters with 8 bits | Jean-Marc Valin | |
Only code non-degenerate symbols, which makes the encoder faster | |||
2023-11-06 | Vectorize DRED quantization | Jean-Marc Valin | |
2023-11-04 | Match silenced overflow checks in the sse4.1 version | Felicia Lim | |
Update silk/x86/NSQ_del_dec_sse4_1.c to match the remaining silk/NSQ_del_dec.c changes made in https://gitlab.xiph.org/xiph/opus/-/commit/c913dc38 Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> | |||
2023-11-03 | Silence some overflow checks | Felicia Lim | |
Co-authored-by: James Zern <jzern@google.com> | |||
2023-11-03 | Vectorizing sgemv for multiples of 4 with SSE | Jean-Marc Valin | |
2023-11-03 | Fix silly bug in CELT Deep PLC | Jean-Marc Valin | |
2023-11-03 | Fix CELT PLC for single packet between losses | Jean-Marc Valin | |
Avoids switching to CNG unless we just have a "refresh packet" | |||
2023-11-02 | updated moc method | Jan Buethe | |
2023-10-30 | Don't try to use models that aren't loaded | Jean-Marc Valin | |
2023-10-30 | Speed up general case for float matrix multiply | Jean-Marc Valin | |
2023-10-30 | Don't use reserved identifiers for include guards | Jean-Marc Valin | |
2023-10-30 | Update blob loading code | Jean-Marc Valin | |
2023-10-29 | Use log approximation when possible | Jean-Marc Valin | |
2023-10-29 | Reusing already-optimized celt_fir() | Jean-Marc Valin | |
2023-10-29 | Silence some warnings | Jean-Marc Valin | |
Including removing useless code | |||
2023-10-29 | Pitch and fargan model updates | Jean-Marc Valin | |
Removing one of the 2d conv layers for pitch estimation reduces complexity without noticeable degradation. FARGAN model has more adversarial training. Also, no need for the double precision in the low-pass filter. | |||
2023-10-27 | Fix CELT deep PLC bugs | Jean-Marc Valin | |
The sinc filter offset was incorrectly handled. Since it perfectly compensates for the analysis offset, nothing has to be done. Also, the preemphasis memory was never initialized. | |||
2023-10-24 | cleanup | Jean-Marc Valin | |
2023-10-24 | Slightly more continuous analysis | Jean-Marc Valin | |
2023-10-21 | Removing more useless code | Jean-Marc Valin | |
2023-10-21 | Stop using tansig_table.h (both copies) | Jean-Marc Valin | |
2023-10-21 | Cleanup | Jean-Marc Valin | |
2023-10-21 | nnet.h no longer needs to #include "vec.h" | Jean-Marc Valin | |
2023-10-20 | more cleanup | Jean-Marc Valin | |
2023-10-20 | Remove unneeded functions in nnet.c | Jean-Marc Valin | |
2023-10-20 | Quantizing pitchdnn and rdovae weights | Jean-Marc Valin | |
2023-10-20 | added 16kHz version of opus_compare in python | Jan Buethe | |
2023-10-20 | finalized quantization option in export_rdovae_weights.py | Jan Buethe | |
2023-10-20 | Doing some unrolling on ARM/Neon | Jean-Marc Valin | |
2023-10-20 | Unroll the 3x3 convolution case | Jean-Marc Valin | |
Gets us about 2x speedup on x86 | |||
2023-10-19 | Marking RDOVAE layers to quantize | Jean-Marc Valin | |
2023-10-19 | prepared quantization implementation for DRED | Jan Buethe | |
2023-10-19 | restructured osce readme | Jan Buethe | |
2023-10-19 | added LACE/NoLACE checkpoint URL | Jan Buethe | |
2023-10-19 | Support OPUS_SET_COMPLEXITY() on decoder side | Jean-Marc Valin | |
Controls whether deep PLC is enabled | |||
2023-10-19 | Avoiding work on the PLC update side | Jean-Marc Valin | |
Shift computation to concealment | |||
2023-10-19 | Simplifying the DRED/PLC code | Jean-Marc Valin | |
2023-10-18 | Don't call the libm tanh() | Jean-Marc Valin | |
2023-10-18 | Making the build possible without the models | Jean-Marc Valin | |
No dependency on the data files if no DNN code enabled |