Functions — github.com/OpenNMT/CTranslate2

CodeHub MCPcopy Create free account

hub / github.com/OpenNMT/CTranslate2 / functions

Functions2,508 in github.com/OpenNMT/CTranslate2

⨍ Functions 2,508◇ Types & classes 601↳ Endpoints 4

MethodScopedMPISetter

src/devices.cc:141

MethodSequenceEncoderReplica

include/ctranslate2/models/language_model.h:116

MethodSequenceGeneratorReplica

include/ctranslate2/models/language_model.h:35

MethodSequenceToSequenceReplica

include/ctranslate2/models/sequence_to_sequence.h:47

MethodSinusoidalPositionEncoder

src/layers/common.cc:231

src/ops/slide.cc:10

src/ops/softmax.cc:12

src/ops/split.cc:10

include/ctranslate2/ops/squeeze.h:12

MethodStorageView

src/storage_view.cc:11

src/ops/sum.cc:7

MethodSuppressSequences

src/decoding_utils.cc:108

MethodSuppressTokens

src/decoding_utils.cc:156

MethodSuppressTokensBegin

src/decoding_utils.cc:172

tests/model_test.cc:7

tests/storage_view_test.cc:4

Merged self+cross attention (T5Gemma2 style): self-attention layer also projects encoder memory through a separate `memory_kv` linear and concatenates

tests/attention_test.cc:129

tests/batching_test.cc:6

tests/decoding_test.cc:5

tests/layers_test.cc:8

tests/ops_test.cc:6

tests/translator_test.cc:381

MQA: All heads share same K/V

tests/attention_test.cc:57

tests/storage_view_test.cc:67

tests/primitives_test.cc:8

tests/layers_test.cc:34

tests/ops_test.cc:128

tests/translator_test.cc:53

FunctionTestMinMax

tests/ops_test.cc:1126

MethodTextLineReader

include/ctranslate2/batch_reader.h:85

MethodThreadPool

src/thread_pool.cc:127

src/ops/tile.cc:8

src/ops/topk.cc:8

src/ops/topp_mask.cc:10

MethodTransformerDecoder

src/layers/transformer.cc:487

MethodTransformerDecoderLayer

src/layers/transformer.cc:150

MethodTransformerEncoder

src/layers/transformer.cc:405

MethodTransformerEncoderLayer

src/layers/transformer.cc:54

MethodTransformerModel

src/models/transformer.cc:36

MethodTranslationResult

include/ctranslate2/translation.h:94

MethodTranspose

src/ops/transpose.cc:8

MethodTranspose

include/ctranslate2/ops/transpose.h:10

MethodUnsqueeze

include/ctranslate2/ops/unsqueeze.h:12

MethodUseTrueFp16GemmInScope

src/cuda/utils.h:120

MethodVectorReader

src/batch_reader.cc:105

MethodVocabulary

src/vocabulary.cc:36

MethodVocabularyMap

src/vocabulary_map.cc:7

MethodWav2Vec2BertEncoder

src/layers/wav2vec2bert.cc:166

MethodWav2Vec2BertReplica

src/models/wav2vec2bert.cc:60

MethodWav2Vec2Encoder

src/layers/wav2vec2.cc:48

MethodWav2Vec2LayerNormConvLayer

src/layers/wav2vec2.cc:7

MethodWav2Vec2PosConvLayer

src/layers/wav2vec2.cc:30

MethodWav2Vec2Replica

src/models/wav2vec2.cc:60

MethodWhisperEncoder

src/layers/whisper.cc:8

MethodWhisperReplica

src/models/whisper.cc:59

Method__aiter__

python/ctranslate2/extensions.py:454

Method__anext__

python/ctranslate2/extensions.py:458

(self, assert_condition, error_message)

python/ctranslate2/converters/utils.py:121

python/ctranslate2/converters/opennmt_tf.py:92

python/ctranslate2/converters/opennmt_tf.py:323

(self, model, tokenizer)

python/ctranslate2/converters/transformers.py:210

(self, *args, **kwargs)

python/ctranslate2/specs/model_spec.py:85

(self, rel_path, runs)

tools/benchmark/benchmark_all.py:33

tools/benchmark/opennmt_ende_wmt14/opennmt_py/convertv2_v3.py:20

tools/benchmark/opennmt_ende_wmt14/opennmt_py/convertv2_v3.py:25

tools/benchmark/opennmt_ende_wmt14/opennmt_py/convertv2_v3.py:30

python/tests/test_spec.py:14

python/tests/test_spec.py:292

(self, process_func, *args, **kwargs)

python/ctranslate2/extensions.py:434

Initializes the OPUS-MT converter. Arguments: model_dir: Path the OPUS-MT model directory.

python/ctranslate2/converters/opus_mt.py:12

Initializes the Marian converter. Arguments: model_path: Path to the Marian model (.npz file). vocab_paths: Paths to the

python/ctranslate2/converters/marian.py:25

python/ctranslate2/converters/utils.py:118

Initializes the Fairseq converter. Arguments: model_path: Path to the Fairseq PyTorch model (.pt file). data_dir: Path to

python/ctranslate2/converters/fairseq.py:107

Initializes the converter. Arguments: model: An initialized and fully-built ``opennmt.models.Model`` instance.

python/ctranslate2/converters/opennmt_tf.py:72

Initializes the OpenAI GPT-2 converter. Arguments: model_dir: Path to the OpenAI GPT-2 model directory.

python/ctranslate2/converters/openai_gpt2.py:12

Initializes the converter. Arguments: model_name_or_path: Name of the pretrained model to download, or path to the dire

python/ctranslate2/converters/transformers.py:70

Initializes the OpenNMT-py converter. Arguments: model_path: Path to the OpenNMT-py PyTorch model (.pt file). unsafe_dese

python/ctranslate2/converters/opennmt_py.py:166

Initializes the OpenNMT-py converter. Arguments: model_path: Path to the OpenNMT-py PyTorch model (.pt file).

python/ctranslate2/converters/eole_ct2.py:154

Initializes a Transformer encoder specification. Args: num_layers: Number of layers. num_heads: Number of attention heads

python/ctranslate2/specs/transformer_spec.py:11

Initializes a Transformer decoder specification. Args: num_layers: Number of layers. num_heads: Number of attention heads

python/ctranslate2/specs/transformer_spec.py:118

( self, relative_position=False, relative_attention_bias=False, ffn_glu=False,

python/ctranslate2/specs/transformer_spec.py:296

( self, with_encoder_attention=True, relative_position=False, relative_attenti

python/ctranslate2/specs/transformer_spec.py:347

python/ctranslate2/specs/transformer_spec.py:453

Initializes the configuration for Transformer models. Args: layer_norm_epsilon: The layer norm epsilon value. **kwargs: A

python/ctranslate2/specs/transformer_spec.py:460

Initializes a Transformer model specification. Args: encoder: The encoder specification. decoder: The decoder specificati

python/ctranslate2/specs/transformer_spec.py:477

Initializes the configuration for Transformer decoder models. Args: layer_norm_epsilon: The layer norm epsilon value. **k

python/ctranslate2/specs/transformer_spec.py:601

Initializes a Transformer decoder model specification. Args: decoder: The decoder specification.

python/ctranslate2/specs/transformer_spec.py:614

Initializes the configuration for Transformer encoder models. Args: layer_norm_epsilon: The layer norm epsilon value. **k

python/ctranslate2/specs/transformer_spec.py:761

Initializes a Transformer encoder model specification. Args: encoder: The encoder specification. pooling_layer: Add the p

python/ctranslate2/specs/transformer_spec.py:774

python/ctranslate2/specs/wav2vec2_spec.py:11

( self, feat_layers, num_layers, num_heads, vocab_size, return

python/ctranslate2/specs/wav2vec2_spec.py:16

python/ctranslate2/specs/wav2vec2_spec.py:49

python/ctranslate2/specs/wav2vec2_spec.py:55

(self, feat_layers, num_layers, num_heads, return_hidden)

python/ctranslate2/specs/wav2vec2_spec.py:60

( self, self_attention=False, relative_position=False, relative_asymmetric_pos

python/ctranslate2/specs/attention_spec.py:18

( self, suppress_ids: Optional[List[int]] = None, suppress_ids_begin: Optional[List[in

python/ctranslate2/specs/whisper_spec.py:11

(self, num_layers, num_heads)

python/ctranslate2/specs/whisper_spec.py:69

Initializes the configuration with a set of parameters.

python/ctranslate2/specs/model_spec.py:295

Initializes the configuration for sequence-to-sequence models. Args: unk_token: The unknown token. bos_token: The start o

python/ctranslate2/specs/model_spec.py:429

Initializes a sequence to sequence model specification.

python/ctranslate2/specs/model_spec.py:467

← previous next →1,001–1,100 of 2,508, ranked by callers