Sync pybind11 with master #4204
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Error message: g++: error: cudafeat/kaldi-*.a: No such file or directory g++: error: cudadecoder/kaldi-*.a: No such file or directory Makefile:67: recipe for target 'biglib' failed make: *** [biglib] Error 1
…-asr#4026) * Add portaudio build instructions. Need to test on a clean kaldi pull. * Correct location of portaudio.props * Small edit to INSTALL.md * Fix portaudio directory * Change paths back to match readme
…di-asr#4023 (kaldi-asr#4033) - tcp-server: - fixing a bug in ivector extractor with weighted frames, - it was occuring during hang-up, the weights were incomplete (not updated...) - and it was failing in this way: ```ASSERTION_FAILED (online2-tcp-nnet3-decode-faster[5.5.675-0185c]:UpdateStatsUntilFrameWeighted():online-ivector-feature.cc:285) Assertion failed: (frame >= 0 && frame < this->NumFramesReady() && delta_weights_provided_ && ! updated_with_no_delta_weights_ && frame <= most_recent_frame_with_weight_)``` - make_bottleneck_features: there was a bug in CMVN pipeline that i recently introduced (sorry about that...)
…t_unk_arpa.pl (kaldi-asr#4053) The --fixed-value parameter of adjsut_unk_arpa.pl was not having the intended effect of "higher order n-grams containing the OOV dict entry remaining untouched". When --fixed-value=true, the unigram entry was getting set correctly, but the higher order n-grams were also being scaled.
…-asr#4073) This reverts commit 401bb1f.
Co-authored-by: Jan 'Yenda' Trmal <jtrmal@apptek.com>
Co-authored-by: jasontu <jasontu@wezhuiyi.com>
…aldi-asr#4083) lattice-copy does not accept --acoustic-scale
* Partial hypotheses * PR comments * Endpointing * PR comments * Neg non-em partial traceback bug fix
…aldi-asr#4133) I could not reproduce the problem with the configuration as close as was reported in https://groups.google.com/d/msg/kaldi-help/5smFF2RW9do/P3Ml5iITAgAJ, with the exception of the compiler packaging version: I used g++ 7.3.1 (Red Hat 7.3.1-6) while the problem was reported against the package version 7.3.1-2. I also could not find similar reports anywhere else. Patching the automake-generaated Makefile is a cure worse than the disease, and I certainly disfavor this solution. It will hit us hard when we upgrade OpenFST. And, since the variable OPENFST_VERSION can be set in make's command line, this solution is certainly asking for a trouble.
…aldi-asr#4168) `make clean` in tools/ fails if openfst/ has not been configured. Let make ignore this failure.
…re/s5 (kaldi-asr#4171) The script has a shebang which explicitly selects python3.
I need to review them without pacing. Some are worth revisiting.
…aldi-asr#4098) - Add the `order` option for all `ngram` commands, so that the script works for higher-order (>3) n-gram LMs - Fix the hard-coded lambda value for LM interpolation
X-Ref: usnistgov/SCTK#24 Close: kaldi-asr#4155 Co-authored-by: Jan 'Yenda' Trmal <jtrmal@apptek.com>
Co-authored-by: Jan 'Yenda' Trmal <jtrmal@apptek.com>
That's what the C++ standard has always required, and g++8 issued a warning about incorrect member construction order.
nnet3-xvector-compute passed the address of an always true bool variable to a formal parameter declared as bool. No functional change here, it worked all the same, only for a wrong reason. The rest is formatting code for readability and code style.
* The link in malach/s5 missed the target script. * Some tuning scripts in swbd/s5c didn't have the x bit set
…i-asr#4191) Clang 10 reports that the copy of the string returned by siter.Symbol() will be freed at the end of the line, so its c_str pointer will become invalid: src/fstext/pre-determinize-inl.h:239:23: error: object backing the pointer will be destroyed at the end of the full-expression [-Werror,-Wdangling-gsl] const char *sym = siter.Symbol().c_str(); ^~~~~~~~~~~~~~
Support scope: * Reading and writing of .npy files. * .npy format versions 2.0 and 3.0. * Little- and big-endian files. * Note that we only handle the case when number of bytes of the HEADER_LEN is 4. Non-ASCII UTF-8 characters are NOT supported.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
for #3797