Age | Commit message (Collapse) | Author | |
---|---|---|---|
2021-04-29 | factor mask | Hieu Hoang | |
2021-04-29 | start factor mask | Hieu Hoang | |
2021-04-29 | start broadcast | Hieu Hoang | |
2021-04-29 | move cache variables into shortlist class | Hieu Hoang | |
2021-04-29 | restart | Hieu Hoang | |
2021-04-28 | Remove unused memoized_ variable (#852) | Kenneth Heafield | |
2021-04-26 | Merged PR 18612: Early stopping on first, all, or any validation metrics | Roman Grundkiewicz | |
Adds `--early-stopping-on first|all|any` allowing to decide if early stopping should take into account only first, all, or any validation metrics. Feature request: https://github.com/marian-nmt/marian-dev/issues/850 Regression tests: https://github.com/marian-nmt/marian-regression-tests/pull/79 | |||
2021-04-21 | update version | Marcin Junczys-Dowmunt | |
2021-04-21 | Merge branch 'master' of github.com:marian-nmt/marian-dev into pmaster | Marcin Junczys-Dowmunt | |
2021-04-20 | fix depth-scaling in FFN | Marcin Junczys-Dowmunt | |
2021-04-14 | Compute tensor size using integers (#851) | Kenneth Heafield | |
2021-04-11 | update version | Marcin Junczys-Dowmunt | |
2021-04-11 | Merge branch 'master' of ↵ | Marcin Junczys-Dowmunt | |
vs-ssh.visualstudio.com:v3/machinetranslation/Marian/marian-dev | |||
2021-04-11 | Merge branch 'pmaster' | Marcin Junczys-Dowmunt | |
2021-04-11 | fix beam_search ABORT when enable openmp and OMP_NUM_THREADS > 1 (#767) | huangjq0617 | |
2021-04-10 | Merged PR 18505: RMSNorm on GPU | Martin Junczys-Dowmunt | |
Support for RMSNorm as drop-in replace for LayerNorm from _Biao Zhang; Rico Sennrich (2019). Root Mean Square Layer Normalization_. Enabled in Transformer model via `--transformer-postprocess dar` instead of `dan`. | |||
2021-04-09 | Merged PR 18531: Install GCC in Azure pipelines | Roman Grundkiewicz | |
This fixes Azure pipelines after recent changes in Azure-hosted runners removing GCC 8 and older on some Ubuntu images. GCC is now installed explicitly via `apt-get`. | |||
2021-04-09 | synced with public master | Marcin Junczys-Dowmunt | |
2021-04-09 | Update VERSION | Marcin Junczys-Dowmunt | |
2021-04-09 | Allow to choose fine-grained CPU intrinsics on as CMake options (#849) | Marcin Junczys-Dowmunt | |
* allow to choose fine-grained CPU intrinsics on as CMake options * inform user that e.g. -DCOMPILE_AVX2=off will be ignored with -march=native if there is compiler support | |||
2021-04-09 | Create VERSION | Marcin Junczys-Dowmunt | |
2021-04-09 | Adds better Affine support for GPUs when using CUDA 11. Introduces a new ↵ | rhenry-nv | |
bias addition kernel for CUDA < 11 (#778) Co-authored-by: Marcin Junczys-Dowmunt <marcinjd@microsoft.com> | |||
2021-04-08 | Fix Ubuntu GitHub checks (#848) | Roman Grundkiewicz | |
* Change ubuntu-latest to ubuntu-18.04 * Install gcc/g++ | |||
2021-04-08 | Revert "remove TC_MALLOC from optional dependencies (#840)" | Marcin Junczys-Dowmunt | |
This reverts commit 096c48e51cd2e61bb275345d7cca99cbfd6bc5c7. | |||
2021-03-31 | Merged PR 18366: Fix generation of special control characters for default ↵ | Rohit Jain | |
vocabulary This PR extends the --allow-special feature to default vocabulary items as well. If the default vocabulary is provided with symbols ostensibly generated from the SentencePiece Byte Fallback mechanism, we suppress the control characters from that list. | |||
2021-03-30 | Update submodule examples | Roman Grundkiewicz | |
2021-03-26 | Merged PR 18309: Cleaner suppression of unwanted output words | Martin Junczys-Dowmunt | |
This PR adds cleaner suppression of unwanted output words. We identified a situation where SPM with byte-fallback can generate random bytes with output-sampling. That is particularly harmful when that random bytes happens to be a newline symbol. Here we suppress newline in output unless explicitly wanted. | |||
2021-03-24 | Merge branch 'pmaster' | Marcin Junczys-Dowmunt | |
2021-03-24 | Properly copy the entire vector in the int16_t case (#845) | Nikolay Bogoychev | |
Fixes #842 #843 #844 | |||
2021-03-23 | Revert "start lsh shortlist" | Hieu Hoang | |
This reverts commit 415769fb2f953c5abf88f2d3498f2a46ea3607d7. | |||
2021-03-23 | start lsh shortlist | Hieu Hoang | |
2021-03-22 | Fix FBGEMM build with gcc 9.3+ (#836) | Young Jin Kim | |
2021-03-22 | Display decoder speed statistics with --stat-freq N (#841) | Marcin Junczys-Dowmunt | |
Display decoder time statistics if requested | |||
2021-03-22 | remove TC_MALLOC from optional dependencies (#840) | Marcin Junczys-Dowmunt | |
There seems to be no benefit from TC_MALLOC any more, hence removing. | |||
2021-03-22 | Merge branch 'master' into pmaster | Marcin Junczys-Dowmunt | |
2021-03-19 | Update VERSION | Roman Grundkiewicz | |
2021-03-19 | Fix model loading on architectures where size_t is 32bits (#825) | Nikolay Bogoychev | |
* fix model loading on architectures where size_t is 32bit * Update the changelog Co-authored-by: Roman Grundkiewicz <rgrundkiewicz@gmail.com> | |||
2021-03-19 | Update VERSION | Roman Grundkiewicz | |
2021-03-19 | Add simple unit tests for binary files (#826) | Marcin Junczys-Dowmunt | |
* unit tests for binary file operations * adjust changelog * Set file_ in TemporaryFile for MSVC Co-authored-by: Roman Grundkiewicz <rgrundkiewicz@gmail.com> | |||
2021-03-19 | Fix broken links to MNIST data sets (#838) | Roman Grundkiewicz | |
2021-03-19 | Merged PR 18232: Update VS CMake builds and scripts | Roman Grundkiewicz | |
This PR updates Windows build using Visual Studio CMake compilation with Ninja. It does not affect standard VS compilation or Windows builds on Azure/GitHub CI. List of changes: - Fixed syntax in the script installing dependencies via vcpkg. - Removed installation of Protobuf (already included as a submodule) and Boost 1.72 (the previous solution no longer works with new vcpkg). - Disabled compilation of marian-server in the default setting due to Boost issues. - Disabled compilation of NCCL in the default setting due to an error (see comment in the code). - Updated vs/README. | |||
2021-03-18 | Update GitHub workflows with Ubuntu+CUDA (#837) | Roman Grundkiewicz | |
2021-03-18 | sync public and internal master | Marcin Junczys-Dowmunt | |
2021-03-18 | increase version and update changelog | Marcin Junczys-Dowmunt | |
2021-03-18 | Merged PR 18185: Support for Microsoft legacy binary shortlist | Martin Junczys-Dowmunt | |
Adds support for Microsoft-internal binary shortlist format. | |||
2021-03-17 | Merged PR 18201: Install Boost in Azure pipelines | Roman Grundkiewicz | |
Installing Boost manually in all workflows, because it has been recently removed from Azure/GitHub hosted runners. This should fix recent failures of Marian CI builds. | |||
2021-03-17 | Install Boost in GitHub workflows (#834) | Roman Grundkiewicz | |
* Install Boost in workflows * Compile CUDA realease for all GPU archs * Convert line endings to LF | |||
2021-03-12 | Update VERSION | Roman Grundkiewicz | |
2021-03-12 | Fix missing float template specialisation for elem::Plus (#822) | Graeme | |
* Fix missing float template specialisation for elem::Plus * Update CHANGELOG.md | |||
2021-03-12 | Fix fallback to default paths in MNIST example (#821) | Graeme | |
If --train-sets or --valid-sets are not provided, the fallback to the hard-coded paths does not occur. This commit requires that these entities have a non-empty value. |