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

github.com/OpenNMT/CTranslate2.git - Unnamed repository; edit this file 'description' to name the repository.
summaryrefslogtreecommitdiff
AgeCommit message (Expand)Author
2022-07-29Bump version to 2.21.1v2.21.1Guillaume Klein
2022-07-29Fix conversion of NLLB models without tokenizer_class in config (#889)Guillaume Klein
2022-07-29Update version of action pypa/gh-action-pypi-publishGuillaume Klein
2022-07-27Bump version to 2.21.0v2.21.0Guillaume Klein
2022-07-27Add support for NLLB multilingual models (#878)Guillaume Klein
2022-07-27Fix FP16 dtype error with return_alternatives + return_attention (#884)Guillaume Klein
2022-07-27Add conversion for Pegasus summarization models (#883)Guillaume Klein
2022-07-21Do not stop decoding when EOS is coming from the user input (#880)Guillaume Klein
2022-07-21Update pybind11 to 2.10.0 (#881)Guillaume Klein
2022-07-21Update cibuildwheel to 2.8.1 (#882)Guillaume Klein
2022-07-19Remove checks for old Intel MKL versions (#875)Guillaume Klein
2022-07-18Add conversion test for OpenNMT-py's transformer_lm (#874)Guillaume Klein
2022-07-18Fix converter relative position propagation in opennmt-py lm converter (#873)Valentin Berkes
2022-07-15Add missing docstring for parameter read_batch_size (#872)Guillaume Klein
2022-07-15Remove non needed register_variable_alias overrideGuillaume Klein
2022-07-15Minor fix in environment variables documentationGuillaume Klein
2022-07-15Fix extra character in documentation fileGuillaume Klein
2022-07-15Fix linear layers fusion without bias (#871)Guillaume Klein
2022-07-14Fix missing processing of linear weights in language models (#869)Guillaume Klein
2022-07-13Keep test parameters close to the test caseGuillaume Klein
2022-07-12Compute d_model from the linear layer output, not from the layer normGuillaume Klein
2022-07-12Do not check model arguments for unsupported Fairseq models (#868)Guillaume Klein
2022-07-08Vectorize GELU operator on CPU (#823)Guillaume Klein
2022-07-06Bump version to 2.20.0v2.20.0Guillaume Klein
2022-07-06Fix application of max_decoding_length in return_alternatives mode (#866)Guillaume Klein
2022-07-06Update PyPI classifiers with supported CUDA versionsGuillaume Klein
2022-07-06Remove <pad> token when converting MarianMT models (#865)Guillaume Klein
2022-07-06Highlight important note in Transformers guideGuillaume Klein
2022-07-05Fix asynchronous result state when a runtime exception is raised (#864)Guillaume Klein
2022-07-04Do not modify config passed to OpenNMTTFConverterV2.from_config (#863)Guillaume Klein
2022-07-04List class members at the top of the class documentation (#861)Guillaume Klein
2022-07-04Update FasterTransformer to 5.0 in benchmark script (#860)Guillaume Klein
2022-06-30Prevent setting new attributes in layer specifications (#859)Guillaume Klein
2022-06-30Vectorize accumulator horizontal reduction (#855)Guillaume Klein
2022-06-30Add option no_repeat_ngram_size to prevent repetitions of ngrams (#854)Guillaume Klein
2022-06-30Remove unnecessary libomp installation command on macOS (#857)Guillaume Klein
2022-06-30Fix conversion of OpenNMT-tf models using static position embeddings (#858)Guillaume Klein
2022-06-30Pin PyTorch version for the macOS Python tests (#856)Guillaume Klein
2022-06-28Update Ubuntu and macOS versions used on GitHub runners (#852)Guillaume Klein
2022-06-23Bump version to 2.19.1v2.19.1Guillaume Klein
2022-06-23Minor change in device_index docstringGuillaume Klein
2022-06-23Add missing final layer norm in OPT model (#850)Guillaume Klein
2022-06-23Check if relative position is used directly from the attention layer (#849)Guillaume Klein
2022-06-21Keep verbose mode enabled for Transformers' tokenizer (#848)Guillaume Klein
2022-06-21Factorize layer creation (#846)Guillaume Klein
2022-06-20Do not pass EmbeddingsSpec modules to the base classesGuillaume Klein
2022-06-20Define float16_t type in public headers (#841)Guillaume Klein
2022-06-20Do not mark relative position as optional when they are required (#844)Guillaume Klein
2022-06-20Remove attribute multiply_by_sqrt_depth that is no longer set (#843)Guillaume Klein
2022-06-20Fix conversion of OpenNMT-tf V1 checkpoints with the new converter (#845)Guillaume Klein
generated by cgit v1.2.3 (git 2.25.1) at 2024-06-06 11:29:05 +0300