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

↓ 6 callersMethodmaybe_t

(self, w)

src/adapters/methods/lora.py:861

↓ 6 callersMethodpad

(self, x, lora, fill_value=None)

src/adapters/methods/lora.py:1182

↓ 6 callersMethodpredict

Run prediction and returns predictions and potential metrics. Depending on the dataset and your use case, your test dataset may cont

examples/pytorch/dependency-parsing/utils_udp.py:242

↓ 6 callersMethodtraining_pass

(self, model)

tests/test_misc/test_adapter_composition.py:66

↓ 5 callersMethod__init__

(self, dim, mlp_ratio)

tests/test_methods/test_all_custom_interfaces.py:67

↓ 5 callersMethodadd_seq2seq_lm_head

Adds a sequence-to-sequence language modeling head on top of the model. Args: head_name (str): The name of the head.

src/adapters/heads/model_mixin.py:518

↓ 5 callersMethodadd_tagging_head

Adds a token classification head on top of the model. Args: head_name (str): The name of the head. num_label

src/adapters/heads/model_mixin.py:438

↓ 5 callersMethodbatched_training_pass

(self, model)

tests/test_misc/test_adapter_composition.py:73

↓ 5 callersMethodcompose_single

Forwards the given state through the given single adapter. Args: adapter_setup (str): The name of the adapter. state

src/adapters/methods/adapter_layer_base.py:339

↓ 5 callersMethodforward

(self, outputs, cls_output=None, attention_mask=None, return_dict=False, **kwargs)

src/adapters/heads/base.py:337

↓ 5 callersMethodget_labels_dict

(self)

src/adapters/model_mixin.py:2633

↓ 5 callersMethodget_output_embeddings

(self)

src/adapters/heads/base.py:111

↓ 5 callersMethodinit_adapters

This method initializes adapter modules and fusion modules from the model config.

src/adapters/model_mixin.py:467

↓ 5 callersMethodpre_block

Optional state pre-processing method which is invoked before passing the state to the first child block of a composition. By default,

src/adapters/methods/adapter_layer_base.py:262

↓ 5 callersMethodrun_seq2seq_quick

( self, distributed=False, extra_args_str=None, predict_with_generate=True,

tests/test_misc/test_adapter_trainer/test_adapter_trainer_ext.py:56

↓ 5 callersMethodset_active_embeddings

Sets the active embedding for the forward pass of the model Args: name: The name of the embedding that should be used

src/adapters/model_mixin.py:371

↓ 5 callersMethodshare_parameters

( self, name: str, adapter_names: List, reference_adapter_name: Optional[str]

src/adapters/methods/lora.py:583

↓ 4 callersMethod__init__

( self, n_layers: int, n_heads: int, input_size: int, config: PrefixTu

src/adapters/methods/prefix_tuning.py:19

↓ 4 callersMethod_get_cls_output

(self, outputs, **kwargs)

src/adapters/heads/base.py:117

↓ 4 callersMethod_store_gating_score

(self, adapter_name, gating_score)

src/adapters/methods/adapter_layer_base.py:64

↓ 4 callersMethodadd_causal_lm_head

Adds a causal language modeling head on top of the model. Args: head_name (str): The name of the head. activ

src/adapters/heads/model_mixin.py:502

↓ 4 callersMethodadd_custom_head

(self, head_type, head_name, overwrite_ok=False, set_active=True, **kwargs)

src/adapters/heads/model_mixin.py:302

↓ 4 callersMethodadd_masked_lm_head

Adds a masked language modeling head on top of the model. Args: head_name (str): The name of the head. activ

src/adapters/heads/model_mixin.py:488

↓ 4 callersMethodbuild_rand_ids_tensor

Creates a random tensor of type torch.long with the given shape with random values in range 0 - (vocab_size-1).

tests/test_methods/base.py:64

↓ 4 callersMethodcom_inv

Inverts the composition operation between existing and injected weights.

src/adapters/methods/lora.py:295

↓ 4 callersFunctiondequantize_bnb_weight

Helper function to dequantize 4bit or 8bit bnb weights. If the weight is not a bnb quantized weight, it will be returned as is.

src/adapters/methods/utils.py:32

↓ 4 callersMethodfilter_func

(self, adapter_name)

src/adapters/loading.py:383

↓ 4 callersMethodfilter_func

(self, adapter_fusion_name)

src/adapters/loading.py:673

↓ 4 callersMethodfrom_dict

(cls, data)

src/adapters/composition.py:65

↓ 4 callersMethodget_dataset

(self, tokenizer=None)

tests/test_methods/base.py:130

↓ 4 callersMethodget_input_samples

(self, shape=None, vocab_size=5000, config=None, **kwargs)

tests/test_methods/base.py:106

↓ 4 callersMethodget_invertible_adapter

(self)

src/adapters/model_mixin.py:140

↓ 4 callersMethodinit_adapters

(self, model_config, adapters_config)

src/adapters/models/t5/mixin_t5.py:77

↓ 4 callersFunctioninit_adapters_config

Initializes the adapters config object of the model to enable adapter support. Also make required changes to the model's config. Args:

src/adapters/wrappers/configuration.py:75

↓ 4 callersMethoditer_layers

(self)

src/adapters/models/t5/mixin_t5.py:103

↓ 4 callersMethodload_weights

( self, save_directory, filter_func, rename_func=None, loading_info=No

src/adapters/loading.py:157

↓ 4 callersMethodload_weights_config

(self, save_directory)

src/adapters/loading.py:102

↓ 4 callersFunctionresolve_path

(path)

src/adapters/utils.py:411

↓ 4 callersMethodsave_weights

(self, save_directory, filter_func)

src/adapters/loading.py:85

↓ 4 callersMethodsave_weights_config

(self, save_directory, config, meta_dict=None)

src/adapters/loading.py:73

↓ 4 callersMethodto_dict

(self)

src/adapters/interface.py:125

↓ 4 callersMethodto_dict

(self)

src/adapters/configuration/adapter_config.py:836

↓ 3 callersMethod__init__

( self, model, weights_name, config_name, use_safetensors: bool = Fals

src/adapters/loading.py:249

↓ 3 callersMethod_check_svd_weights

(self, delta_w, merged_lora, svd_rank, atol=1e-5)

tests/test_methods/method_test_impl/peft/test_dora.py:115

↓ 3 callersMethod_check_svd_weights

(self, delta_w, merged_lora, svd_rank, atol=1e-5)

tests/test_methods/method_test_impl/peft/test_lora.py:107

↓ 3 callersMethod_get_used_heads

(self, head_name: str = None)

src/adapters/heads/model_mixin.py:550

↓ 3 callersMethodadapter_to

Moves the adapter with the given name to the specified device and data type. Args: name (str): The name of the adapter t

src/adapters/model_mixin.py:1514

↓ 3 callersMethodadd

Adds a new adapter of the name to the model config. Args: adapter_name (str): The name of the adapter. confi

src/adapters/configuration/model_adapters_config.py:112

↓ 3 callersMethodadd_invertible_adapter

Adds an invertible adapter module for the adapter with the given name. If the given adapter does not specify an invertible adapter co

src/adapters/model_mixin.py:69

↓ 3 callersMethodadd_prediction_head_from_config

( self, head_name: str, config: dict, overwrite_ok: bool = False, set_

src/adapters/heads/model_mixin.py:201

↓ 3 callersMethodbuild

(self, model)

src/adapters/heads/language_modeling.py:41

↓ 3 callersMethodbuild_rand_tensor

Creates a random tensor of the given shape.

tests/test_methods/base.py:57

↓ 3 callersMethodcompose

The main composition forward method which recursively calls the composition blocks forward methods. This method should be called by the forwar

src/adapters/methods/adapter_layer_base.py:583

↓ 3 callersMethoddelete_adapter

Deletes the adapter with the specified name from the model. Args: adapter_name (str): The name of the adapter.

src/adapters/model_mixin.py:954

↓ 3 callersMethoddelete_adapter_fusion

Deletes the AdapterFusion layer of the specified adapters. Args: adapter_names (Union[Fuse, list, str]): AdapterFusion l

src/adapters/model_mixin.py:977

↓ 3 callersMethodget_adapter

Returns a dictionary with all weights of the adapter with the specified name. Args: name (str): The adapter name.

src/adapters/model_mixin.py:1472

↓ 3 callersFunctionget_head_config_and_rename_list

(model_class_name, head_name, label2id, num_labels=None, return_rename_func=True)

src/adapters/head_utils.py:802

↓ 3 callersMethodget_input_samples

(self, shape, vocab_size=5000)

tests/test_misc/test_adapter_safetensors.py:24

↓ 3 callersMethodinit_adapters

(self, model_config, adapters_config)

src/adapters/models/bert/mixin_bert.py:61

↓ 3 callersMethodinvertible_adapters

(self)

src/adapters/model_mixin.py:192

↓ 3 callersFunctionis_path_in_dir

(path, base_dir)

src/adapters/utils.py:415

↓ 3 callersMethodlast

(self)

src/adapters/composition.py:36

↓ 3 callersMethodload_embeddings

Load a saved embedding from the given path. If the embedding was saved with a tokenizer it is returned Args: path: the p

src/adapters/model_mixin.py:251

↓ 3 callersMethodload_weights_from_state_dict

( self, state_dict, filter_func, rename_func=None, loading_info=None,

src/adapters/loading.py:190

↓ 3 callersFunctionmain

()

examples/pytorch/translation/run_translation.py:263

↓ 3 callersFunctionresolve_adapter_config

Resolves a given adapter configuration specifier to a full configuration dictionary. Args: config (Union[dict, str]): The configurat

src/adapters/utils.py:548

↓ 3 callersFunctionresolve_adapter_path

Resolves the path to a pre-trained adapter module. Note: If attempting to resolve an adapter from the Hub, adapter_config and model_name must

src/adapters/utils.py:741

↓ 3 callersMethodrun_generate_test

(self, adapter_config, max_new_tokens=32)

tests/test_methods/method_test_impl/base.py:458

↓ 3 callersMethodrun_parallel_training_test

(self, adapter_config, filter_key)

tests/test_methods/method_test_impl/composition/test_parallel.py:175

↓ 3 callersMethodrun_trainer

( self, max_len: int, model_name: str, num_train_epochs: int, learning

tests/test_misc/test_adapter_trainer/test_adapter_trainer_ext.py:255

↓ 3 callersMethodsave_all_adapters

Saves all adapters of this model together with their configuration to subfolders of the given location. Args: save_direc

src/adapters/model_mixin.py:1340

↓ 3 callersMethodsave_embeddings

Saves the embedding with the given name. If a tokenizer is passed as well the tokenizer is saved together with the embedding.

src/adapters/model_mixin.py:351

↓ 3 callersMethodtrainings_run

( self, model, lr=1e-5, steps=8, batch_size=2, gradient_accumu

tests/test_methods/method_test_impl/composition/test_multi_task.py:42

↓ 3 callersMethodvslice

Slices the given state along the batch size (vertical) dimension. This is e.g. used by the BatchSplit and Parallel composition blocks. IMPORTA

src/adapters/methods/adapter_layer_base.py:282

↓ 2 callersMethod__init__

( self, in_dim: int, r_dim: int, orthogonal: bool = False, subtract_pr

src/adapters/methods/reft.py:18

↓ 2 callersMethod__init__

( self, model, head_name, vocab_size=None, embedding_size=None,

src/adapters/heads/language_modeling.py:13

↓ 2 callersMethod__init__

(self)

src/adapters/models/t5/mixin_t5.py:56

↓ 2 callersMethod__setattr__

(self, name, value)

src/adapters/configuration/adapter_config.py:264

↓ 2 callersMethod_add_adapter_weights

Helper method that performs the actual parameter additions when adding a new adapter.

src/adapters/model_mixin.py:730

↓ 2 callersMethod_add_tied_weights_keys

Internal method to add adapter-specific keys to the list of tied weights keys.

src/adapters/model_mixin.py:436

↓ 2 callersMethod_assert_adapter_unavailable

Check wether the adapter name is not present in the model's adapter config and has not been created.

tests/test_methods/method_test_impl/base.py:30

↓ 2 callersMethod_average_invertible_adapter

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

src/adapters/model_mixin.py:106

↓ 2 callersMethod_bsz

Returns the batch size of the given state.

src/adapters/methods/adapter_layer_base.py:256

↓ 2 callersMethod_calc_total_dim

(self, shape)

tests/test_methods/base.py:70

↓ 2 callersMethod_copy_hooks_from

(self, module: nn.Module)

src/adapters/methods/lora.py:728

↓ 2 callersMethod_fix_legacy_config

(self, adapter_name, missing_keys)

src/adapters/loading.py:432

↓ 2 callersMethod_get_config_class

Returns the matching config class for the given config dict based on its "architecture" key.

src/adapters/configuration/adapter_config.py:76

↓ 2 callersMethod_get_dropout_prob

(self, model_config)

src/adapters/heads/base.py:68

↓ 2 callersMethod_get_filter_keys

(self, filter_keys, task_names, name)

tests/test_methods/method_test_impl/composition/test_multi_task.py:39

↓ 2 callersMethod_init_mapping

(self)

src/adapters/methods/adapter_layer_base.py:237

↓ 2 callersMethod_init_model_for_train_run

(self, trained_adapter_name, frozen_adapter_name, adapter_config=None)

tests/test_methods/base.py:84

↓ 2 callersMethod_load_adapter_setup_config

( self, load_directory: str )

src/adapters/model_mixin.py:1189

↓ 2 callersMethod_load_adapter_setup_weights

( self, load_directory: str, adapter_setup: AdapterCompositionBlock, custom_we

src/adapters/model_mixin.py:1237

↓ 2 callersMethod_pre_average_adapter_checks

( self, adapter_name: str, adapter_list: List[str], combine_strategy: str,

src/adapters/model_mixin.py:1649

↓ 2 callersMethod_prediction_loop

Prediction/evaluation loop, shared by :obj:`Trainer.evaluate()` and :obj:`Trainer.predict()`. Works both with or without labels.

examples/pytorch/dependency-parsing/utils_udp.py:297

↓ 2 callersFunction_reregister_can_record_outputs

Re-register _can_record_outputs in _CAN_RECORD_REGISTRY after dynamic class replacement. HuggingFace's @check_model_inputs decorator looks up _CA

src/adapters/wrappers/model.py:44

↓ 2 callersMethod_save

(self, output_dir: Optional[str] = None, state_dict=None)

src/adapters/trainer.py:138

↓ 2 callersMethod_save_adapter_card

( self, save_directory: str, adapter_name: str, adapter_repo_name: str,

src/adapters/hub_mixin.py:61

↓ 2 callersMethod_save_adapter_setup_config

( self, save_directory: str, adapter_setup: AdapterCompositionBlock, head_setu

src/adapters/model_mixin.py:1175

↓ 2 callersMethod_save_adapter_setup_weights

( self, save_directory: str, adapter_setup: AdapterCompositionBlock, meta_dict

src/adapters/model_mixin.py:1200

← previous next →101–200 of 1,559, ranked by callers