marian-nmt/marian
1
1
Fork 0
mirror of https://github.com/marian-nmt/marian.git synced 2024-09-17 09:47:34 +03:00
Code Issues Packages Projects Releases Wiki Activity

Update CHANGELOG

Browse Source
This commit is contained in:
Roman Grundkiewicz 2018-01-10 10:59:53 +00:00
parent 379dd9071e
commit 2c26cb1a41
1 changed files with 16 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

26
CHANGELOG.md
View File

@ -7,26 +7,33 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
## [Unreleased]
### Added
- CMake option `-DCOMPILE_SERVER=ON`
### Fixed
- Fixed marian-server compilation with Boost 1.66
## [1.1.3] - 2017-12-06
### Added
- Added back gradient-dropping
### Fixed
- Fixed parameters initializaiton for `--tied-embeddings` during translaton
- Fixed parameters initialization for `--tied-embeddings` during translation
## [1.1.2] - 2017-12-05
### Fixed
- Fixed ensembling with language model and batched decoding
- Fixed attention reduction kernel with large matrices (added missing syncthreads()),
which should fix stability with large batches and beam-size during batched decoding.
- Fixed attention reduction kernel with large matrices (added missing
`syncthreads()`), which should fix stability with large batches and beam-size
during batched decoding.
## [1.1.1] - 2017-11-30
### Added
- Option --max-length-crop to be used together with --max-length N to crop
sentences to length N rather than omitting them.
- Option `--max-length-crop` to be used together with `--max-length N` to crop
sentences to length N rather than omitting them.
- Experimental model with convolution over input characters
### Fixed
@ -53,8 +60,8 @@ sentences to length N rather than omitting them.
- Cyclic learning rate warmup
- More options for learning rate decay, including: optimizer history reset,
repeated warmup
- Continuous inverted square root decay of learning (`--lr-decay-inv-sqrt`) rate
based on number of updates
- Continuous inverted square root decay of learning (`--lr-decay-inv-sqrt`)
rate based on number of updates
- Exposed optimizer parameters (e.g. momentum etc. for Adam)
- Version of deep RNN-based models compatible with Nematus (`--type nematus`)
- Synchronous SGD training for multi-gpu (enable with `--sync-sgd`)
@ -62,7 +69,7 @@ sentences to length N rather than omitting them.
currently only available through the C++ API
- Option `--quiet` to suppress output to stderr
- Option to choose different variants of optimization criterion: mean
cross-entropy, perplexity, cross-entopry sum
cross-entropy, perplexity, cross-entropy sum
- In-process translation for validation, uses the same memory as training
- Label Smoothing
- CHANGELOG.md
@ -72,8 +79,7 @@ sentences to length N rather than omitting them.
### Changed
- Changed shape organization to follow numpy.
- Changed option `--moving-average` to `--exponential-smoothing` and inverted
formula to `s_t = (1 - \alpha) * s_{t-1} + \alpha * x_t`, `\alpha` is now
- Changed option `--moving-average` to `--exponential-smoothing` and inverted formula to `s_t = (1 - \alpha) * s_{t-1} + \alpha * x_t`, `\alpha` is now
`1-e4` by default
- Got rid of thrust for compile-time mathematical expressions
- Changed boolean option `--normalize` to `--normalize [arg=1] (=0)`. New