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

github.com/OpenNMT/OpenNMT-tf.git - Unnamed repository; edit this file 'description' to name the repository.
summaryrefslogtreecommitdiff
AgeCommit message (Expand)Author
2022-11-07Fix error when using guided alignment with mixed precision (#982)HEADmasterGuillaume Klein
2022-10-10Remove recording of the global gradient norm (#978)Guillaume Klein
2022-10-10Add flag to pad each batch to the length bucket boundary (#977)Guillaume Klein
2022-10-03Bump version to 2.29.1v2.29.1Guillaume Klein
2022-09-30Do not consider parameter decay_step_duration as a V1 only parameter (#976)Guillaume Klein
2022-09-29Remove non needed FP32 cast before softmax (#975)Guillaume Klein
2022-09-28Fix count_lines function whe the file is gzipped (#974)Guillaume Klein
2022-09-26Bump version to 2.29.0v2.29.0Guillaume Klein
2022-09-26Remove unnecessary inheritance from object classGuillaume Klein
2022-09-19Add Transformer configurations used in the Scaling NMT paper (#972)Guillaume Klein
2022-09-16Add parameter to automatically share embeddings when possible (#971)Guillaume Klein
2022-09-15Accept list of checkpoints in average_checkpoints method (#969)Guillaume Klein
2022-09-14Fix broken testsGuillaume Klein
2022-09-14Support TensorFlow 2.10 (#968)Guillaume Klein
2022-09-14Use new API to resolve variable names in checkpoints (#967)Guillaume Klein
2022-09-14Use minimal gradient accumulation in batch size autotuning (#966)Guillaume Klein
2022-08-03Use nullcontext instead of suppressGuillaume Klein
2022-07-29Update version for action pypa/gh-action-pypi-publishGuillaume Klein
2022-07-29Bump version to 2.28.0v2.28.0Guillaume Klein
2022-07-29Pad dynamically when preparing the dataset for batch size autotuning (#962)Guillaume Klein
2022-07-29Fix deprecation warnings for using distutils.version.LooseVersion (#963)Guillaume Klein
2022-07-28Make initial learning rate of InvSqrtDecay configurable (#961)Guillaume Klein
2022-07-28Fix link to anchor in tokenization.mdGuillaume Klein
2022-07-28Avoid building batches with more tokens than the user batch size (#959)Guillaume Klein
2022-07-28Add log when the averaged checkpoint is saved (#958)Guillaume Klein
2022-07-28Include sacrebleu==2.2.* in requirements (#960)Guillaume Klein
2022-07-15Add argument to disable bias terms in the multi-head attention (#954)Guillaume Klein
2022-07-15Add argument to disable the bias in the output linear layer (#955)Guillaume Klein
2022-06-28Fix dtype of SequenceRecord's length feature (#953)Guillaume Klein
2022-06-02Bump version to 2.27.1v2.27.1Guillaume Klein
2022-06-02Fix evaluation and scoring with language models (#950)Guillaume Klein
2022-05-30Bump version to 2.27.0v2.27.0Guillaume Klein
2022-05-30Fix inference features returned by the language model inputter (#948)Guillaume Klein
2022-05-30Use simpler structure for the generated API reference (#949)Guillaume Klein
2022-05-27Remove support for older TensorFlow versions 2.4 & 2.5 (#946)Guillaume Klein
2022-05-25Set a minimum elapsed time of 1 ms for logging metrics (#947)Guillaume Klein
2022-05-24Integrate the new CTranslate2 converter for OpenNMT-tf (#945)Guillaume Klein
2022-05-23Support TensorFlow 2.9 (#944)Guillaume Klein
2022-05-16Fix training batch size documentationGuillaume Klein
2022-05-02Fix property formatting in docsGuillaume Klein
2022-04-11Update docs requirements (#940)Guillaume Klein
2022-04-11Rollback to tf.nn.embedding_lookup in non TensorFlow Lite mode (#942)Guillaume Klein
2022-03-31Bump version to 2.26.1v2.26.1Guillaume Klein
2022-03-31Fix documentation build (#938)Guillaume Klein
2022-03-31Bump version to 2.26.0v2.26.0Guillaume Klein
2022-03-16Clarify the inference buffering log (#935)Guillaume Klein
2022-03-16Enable CTranslate2 conversion for models with GELU or Swish activations (#936)Guillaume Klein
2022-03-16Fix WordNoiser on inputs with ranks > 2 (#934)Guillaume Klein
2022-03-09Add learning rate schedule InvSqrtDecay (#929)Guillaume Klein
2022-03-03Use PyPI API token for authentificationGuillaume Klein