Functions — github.com/adapter-hub/adapters

hub / github.com/adapter-hub/adapters / functions

Functions1,559 in github.com/adapter-hub/adapters

⨍ Functions 1,559◇ Types & classes 451↳ Endpoints 1

↓ 1 callersFunction_get_matching_version

(config_entry, org)

src/adapters/utils.py:678

↓ 1 callersMethod_get_save_kwargs

(self)

src/adapters/composition.py:49

↓ 1 callersMethod_init_W

(self, W_left=None, W_right=None, W=None)

src/adapters/methods/modeling.py:672

↓ 1 callersMethod_init_adapters_submodules

(self, model_config, adapters_config)

src/adapters/model_mixin.py:449

↓ 1 callersMethod_init_prompt_embedding

(self, base_model_embeddings: nn.Module)

src/adapters/methods/prompt_tuning.py:76

↓ 1 callersMethod_link_prefix_to_pool

(self, layer)

src/adapters/model_mixin.py:432

↓ 1 callersMethod_load

(cls, path_or_repo_id: str, **kwargs)

src/adapters/interface.py:157

↓ 1 callersMethod_load_adapter_fusions

(self, resume_from_checkpoint)

src/adapters/trainer.py:208

↓ 1 callersMethod_load_adapters

(self, resume_from_checkpoint)

src/adapters/trainer.py:197

↓ 1 callersMethod_load_heads

(self, resume_from_checkpoint)

src/adapters/trainer.py:214

↓ 1 callersMethod_load_module_state_dict

(module, state_dict, start_prefix="")

src/adapters/loading.py:121

↓ 1 callersMethod_merge_subword_tokens

(self, subword_outputs, word_starts)

src/adapters/heads/dependency_parsing.py:134

↓ 1 callersFunction_minimize_dict

(d)

src/adapters/utils.py:139

↓ 1 callersFunction_regex_list_rename_func

(k, rename_list)

src/adapters/head_utils.py:794

↓ 1 callersMethod_rename_legacy_weights

(self, k)

src/adapters/loading.py:406

↓ 1 callersMethod_run_gradient_checkpointing_test_helper

Test that gradient checkpointing produces the same results as normal training Args: adapter_setup_fn: Function that takes

tests/test_methods/method_test_impl/base.py:400

↓ 1 callersMethod_scatter_adapted_states

(self, hidden_states: torch.Tensor, adapted_states: List[torch.Tensor])

src/adapters/methods/reft.py:148

↓ 1 callersMethod_set_layer_hook_for_parallel

(self, layer: nn.Module)

src/adapters/models/albert/mixin_albert.py:60

↓ 1 callersMethod_set_layer_hook_for_parallel

(self, layer: nn.Module)

src/adapters/models/xmod/mixin_xmod.py:31

↓ 1 callersMethod_set_layer_hook_for_parallel

(self, layer: nn.Module)

src/adapters/models/clip/mixin_clip.py:57

↓ 1 callersMethod_set_layer_hook_for_parallel

(self, layer: nn.Module)

src/adapters/models/bert/mixin_bert.py:87

↓ 1 callersMethod_shape

(self, tensor: torch.Tensor, seq_len: int, bsz: int)

src/adapters/models/plbart/modeling_plbart.py:47

↓ 1 callersFunction_split_identifier

(identifier)

src/adapters/utils.py:598

↓ 1 callersMethod_store_fusion_attentions

(self, fusion_name, attentions)

src/adapters/methods/adapter_layer_base.py:81

↓ 1 callersFunction_validate_interface_values

Validates that all values specified in the interface exist in the model. Args: base_model: The base model to validate against

src/adapters/wrappers/model.py:216

↓ 1 callersMethodadapter_fusion_to

Moves the adapter fusion layer with the given name to the specified device and data type. Args: adapter_names (Union[Fus

src/adapters/model_mixin.py:1532

↓ 1 callersMethodadapter_summary

Returns a string summary of all adapters currently added to the model. Each entry in the summary table has the following attributes:

src/adapters/model_mixin.py:1553

↓ 1 callersMethodadd

Parameters ---------- predicted_indices : ``torch.Tensor``, required. A tensor of head index predictions of shape

examples/pytorch/dependency-parsing/utils_udp.py:122

↓ 1 callersMethodadd_adapter

Adds a new adapter module to the layer. Args: adapter_name (str): The name of the new adapter to add. layer_idx (int)

src/adapters/methods/adapter_layer_base.py:90

↓ 1 callersMethodadd_adapter

(self, adapter_name: str, layer_idx: int)

src/adapters/methods/prefix_tuning.py:351

↓ 1 callersMethodadd_adapter

(self, adapter_name: str, layer_idx: int)

src/adapters/methods/invertible.py:22

↓ 1 callersMethodadd_context_args_in_signature

(cls, f)

src/adapters/context.py:130

↓ 1 callersMethodadd_head

Adds a dummy head to the model.

tests/test_methods/base.py:25

↓ 1 callersMethodadd_multiple_choice_head

Adds a multiple choice head on top of the model. Args: head_name (str): The name of the head. num_choices (i

src/adapters/heads/model_mixin.py:414

↓ 1 callersMethodadd_qa_head

Adds a question answering head on top of the model. Args: head_name (str): The name of the head. num_labels

src/adapters/heads/model_mixin.py:455

↓ 1 callersFunctionadjust_length_to_model

(length, max_sequence_length)

examples/pytorch/text-generation/run_generation.py:152

↓ 1 callersMethodapply_weights

(self, hidden_states: torch.Tensor)

src/adapters/methods/lora.py:203

↓ 1 callersMethodattach_labels

(self, inputs)

tests/test_methods/base.py:150

↓ 1 callersMethodattach_labels

(self, inputs)

tests/test_methods/test_on_bert_generation.py:74

↓ 1 callersMethodaverage_head

Adds a new prediction head as a weighted average of a set of existing prediction heads. Args: head_name (str): The name

src/adapters/model_mixin.py:2206

↓ 1 callersMethodaverage_prefix

( self, prefix_name: str, input_adapters: Dict[str, float], combine_strategy:

src/adapters/methods/prefix_tuning.py:215

↓ 1 callersMethodbuild

(self, model)

src/adapters/heads/dependency_parsing.py:79

↓ 1 callersMethodcommon_config_value

Checks whether all adapters in a list share the same config setting for a given attribute and returns the shared value. Args

src/adapters/configuration/model_adapters_config.py:198

↓ 1 callersFunctioncompare_lists_close

Reimplementation of `allclose()` for lists.

utils/back_comp/Utils.py:467

↓ 1 callersMethodcompose_batch_split

For splitting to multiple adapters along the batch size dimension.

src/adapters/methods/adapter_layer_base.py:422

↓ 1 callersFunctioncompute_dora_add_com_inv

This function computes the inverse of the DoRA composition operation i.e the `compute_dora_deltaw` function with the equation w_0' = m * (w_0

src/adapters/methods/lora.py:495

↓ 1 callersFunctioncompute_dora_deltaw

This function calculates the dora update. In the paper, the dora update delta_w is calculated as follows: m * (w_0x + BAx) / ||w_0 + BA||c

src/adapters/methods/lora.py:464

↓ 1 callersFunctionconvert_xmod_checkpoint

(model_name: str, output_dir: str)

utils/convert_xmod_checkpoint.py:15

↓ 1 callersMethodcopy_from

(self, module: nn.Linear)

src/adapters/methods/lora.py:850

↓ 1 callersFunctioncreate_common_voice

Creates a small abstract dataset of 10 samples from the common voice dataset in english.

tests/fixtures/audio_datasets/prepare_audio_datasets.py:9

↓ 1 callersFunctioncreate_common_voice_encoded

Preprocesses the common voice dataset and creates a new encoded version ready for training.

tests/fixtures/audio_datasets/prepare_audio_datasets.py:25

↓ 1 callersFunctioncreate_model

Creates and returns an instance of a specified test model. Args: model_name (str): Specifies which model to instantiate. Raises:

utils/back_comp/Utils.py:142

↓ 1 callersFunctioncreate_speech_commands

Creates a small abstract dataset of 10 samples from the speech commands dataset.

tests/fixtures/audio_datasets/prepare_audio_datasets.py:91

↓ 1 callersFunctioncreate_speech_commands_encoded

Preprocesses the speech commands dataset and creates a new encoded version ready for training.

tests/fixtures/audio_datasets/prepare_audio_datasets.py:109

↓ 1 callersMethodcreate_twin_models

(self, config, adapter_interface, hf_auto_model_class)

tests/test_misc/test_custom_interface_compat.py:74

↓ 1 callersMethodcreate_twin_models

(self)

tests/test_methods/method_test_impl/base.py:202

↓ 1 callersMethodcreate_twin_models

(self)

tests/test_methods/method_test_impl/core/test_adapter_backward_compability.py:12

↓ 1 callersFunctiondecode_tuple

Reconstructs a potentially nested tuple of type `torch.Tensor` as a nested list. Args: tuple_to_decode (tuple): The tuple to decode. R

utils/back_comp/Utils.py:393

↓ 1 callersMethoddelete_adapter

Deletes an adapter module from the layer. Args: adapter_name (str): The name of the adapter to delete.

src/adapters/methods/adapter_layer_base.py:148

↓ 1 callersMethoddelete_adapter

(self, adapter_name: str)

src/adapters/methods/prefix_tuning.py:421

↓ 1 callersMethoddelete_fusion_layer

(self, adapter_names: Union[List, str])

src/adapters/methods/bottleneck.py:134

↓ 1 callersMethoddelete_prefix

(self, prefix_name: str)

src/adapters/methods/prefix_tuning.py:243

↓ 1 callersFunctiondeps_list

(*pkgs)

setup.py:82

↓ 1 callersFunctiondora_merge

This function calculates the weights required to merge with the original pretrained weight matrices.

src/adapters/methods/lora.py:482

↓ 1 callersMethode

(self, s)

src/adapters/methods/modeling.py:546

↓ 1 callersMethodeject_prefix_tuning

Converts the prefix tuning with the given name from the reparameterized form into the flat form. Args: name (str): The n

src/adapters/model_mixin.py:1817

↓ 1 callersMethodenable_prefix

(self, prefix_name: str)

src/adapters/methods/prefix_tuning.py:247

↓ 1 callersMethodextract_input_ids

(self, inputs)

tests/test_methods/base.py:235

↓ 1 callersFunctionfilter_parameters

(model, filter_string)

tests/test_methods/method_test_impl/embeddings/test_adapter_embeddings.py:10

↓ 1 callersFunctionfilter_parameters

(model, filter_string)

tests/test_methods/method_test_impl/composition/test_parallel.py:20

↓ 1 callersFunctionfilter_safe_tarinfos

(members)

src/adapters/utils.py:438

↓ 1 callersFunctionfilter_safe_zipinfos

(members)

src/adapters/utils.py:424

↓ 1 callersMethodforward

(self, embedded_input)

src/adapters/methods/prompt_tuning.py:113

↓ 1 callersMethodforward

(self, hidden_states: Optional[torch.Tensor], layer_input: torch.Tensor)

src/adapters/methods/lora.py:299

↓ 1 callersMethodforward

( self, hidden_states: torch.Tensor, head_mask: Optional[torch.Tensor] = None,

src/adapters/models/beit/modeling_beit.py:197

↓ 1 callersMethodforward

(self, hidden_states: torch.Tensor, input_tensor: torch.Tensor)

src/adapters/models/roberta/modeling_roberta.py:283

↓ 1 callersMethodforward

( self, hidden_states: torch.Tensor, attention_mask: Optional[torch.FloatTensor] = Non

src/adapters/models/albert/modeling_albert.py:177

↓ 1 callersMethodforward

( self, hidden_states: torch.Tensor, head_mask: Optional[torch.Tensor] = None,

src/adapters/models/vit/modeling_vit.py:116

↓ 1 callersMethodforward

(self, hidden_states: torch.Tensor, input_tensor: torch.Tensor)

src/adapters/models/bert/modeling_bert.py:274

↓ 1 callersMethodforward

( self, input_ids: torch.LongTensor = None, encoder_hidden_states: Optional[torch.FloatTensor] = None,

src/adapters/models/whisper/mixin_whisper.py:76

↓ 1 callersMethodforward

(self, hidden_states: torch.Tensor, input_tensor: torch.Tensor)

src/adapters/models/xlm_roberta/modeling_xlm_roberta.py:283

↓ 1 callersMethodforward

Parameters: x: torch.tensor(bs, seq_length, dim) attn_mask: torch.tensor(bs, seq_length) Returns:

src/adapters/models/distilbert/modeling_distilbert.py:282

↓ 1 callersMethodforward

( self, *args, input_ids: Optional[torch.LongTensor] = None, start_positions:

src/adapters/models/t5/mixin_t5.py:128

↓ 1 callersMethodforward_context

This method is called by the ``ForwardContext`` at the beginning of the forward pass.

src/adapters/model_mixin.py:1410

↓ 1 callersMethodfreeze_adapter

(self, adapter_name: str, freeze: bool = True)

src/adapters/methods/prefix_tuning.py:441

↓ 1 callersFunctiongenerate_dummy_data

Generates dummy data for text and vision transformers. Args: model (str, optional): The name of the transformer model. Defaults to an empt

utils/back_comp/Utils.py:350

↓ 1 callersFunctiongenerate_input

(_question, _context)

examples/pytorch/question-answering/run_seq2seq_qa.py:480

↓ 1 callersMethodgenerate_test_methods

Dynamically create test methods for each adapter config.

tests/test_methods/method_test_impl/composition/test_multi_task.py:273

↓ 1 callersMethodget_adapter_config

(self)

tests/test_misc/test_adapter_composition.py:43

↓ 1 callersMethodget_adapter_fusion

(self, adapter_names: Union[List, str])

src/adapters/methods/bottleneck.py:169

↓ 1 callersFunctionget_adapter_interface

(model_name)

src/adapters/wrappers/interfaces.py:133

↓ 1 callersMethodget_context_adapter_setup

(cls)

src/adapters/context.py:58

↓ 1 callersMethodget_dataset

(self, task_type: str = "seq2seq_lm", **kwargs)

tests/test_methods/base.py:228

↓ 1 callersMethodget_dataset_non_batched

(self, config)

tests/test_methods/base.py:142

↓ 1 callersMethodget_dataset_with_task_ids

(self, tasks)

tests/test_methods/method_test_impl/composition/test_multi_task.py:22

↓ 1 callersFunctionget_default

(num_label)

tests/test_misc/test_adapter_save_id2label.py:10

↓ 1 callersMethodget_encoder

(self)

src/adapters/models/t5/adapter_model.py:56

↓ 1 callersFunctionget_from_cache

Given a URL, look for the corresponding file in the local cache. If it's not there, download it. Then return the path to the cached file.

src/adapters/utils.py:242

↓ 1 callersMethodget_from_config

Get the adapter type from a given adapter config. Args: config: The adapter config. Returns: List[s

src/adapters/interface.py:33

↓ 1 callersMethodget_fusion_regularization_loss

(self)

src/adapters/model_mixin.py:1455

← previous next →301–400 of 1,559, ranked by callers