Functions — github.com/VITA-Group/WeLore

CodeHub MCPcopy Create free account

hub / github.com/VITA-Group/WeLore / functions

Functions139 in github.com/VITA-Group/WeLore

⨍ Functions 139◇ Types & classes 21 ↳ Endpoints 1

r""" Args: labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): Labels for computing

peft_pretraining/modeling_llama.py:633

r""" labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*): Labels for computing the sequence classification/regression

peft_pretraining/modeling_llama.py:794

Methodget_decoder

peft_pretraining/modeling_llama.py:628

Functionget_grad_singular_values

lib/rank_utils.py:116

Methodget_input_embeddings

peft_pretraining/modeling_llama.py:613

Methodget_input_embeddings

peft_pretraining/modeling_llama.py:787

Functionget_loaders

(name, nsamples=128, seed=0, seqlen=2048, tokenizer=None)

lib/rank_reduction.py:10

Functionget_loaders

(name, nsamples=128, seed=0, seqlen=2048, tokenizer=None)

lib/rank_utils.py:11

Methodget_output_embeddings

peft_pretraining/modeling_llama.py:619

Functionget_scheculer

( optimizer, *, scheduler_type, num_training_steps, warmup_steps, min_lr_ratio, cy

peft_pretraining/training_utils.py:9

Functionget_singular_values

lib/rank_utils.py:99

Functionmagnitude_pruning

Performs magnitude pruning dimensionality reduction. Only reduces the inner dimensionality, does not affect the shape of the tensor

peft_pretraining/training_utils.py:115

Functionmax_train_tokens_to_number

(max_train_tokens)

peft_pretraining/training_utils.py:214

Functionprepare_calibration_input

(model, dataloader, device)

lib/rank_utils.py:41

Methodprepare_inputs_for_generation

( self, input_ids, past_key_values=None, attention_mask=None, inputs_embeds=None, **kwargs )

peft_pretraining/modeling_llama.py:722

Functionpreprocess_batch

Tokenizing a batch

lib/downstream_utils.py:354

Functionpreprocess_batched

welore_continual_finetune.py:257

Functionrandom_pruning

Performs random pruning dimensionality reduction. Only reduces the inner dimensionality, does not affect the shape of the tensor

peft_pretraining/training_utils.py:104

Functionrank_reduction_dynamic_pruning

(args, model, device, file_name)

lib/rank_utils.py:284

Functionrank_reduction_weight

(args, model, tokenizer, rank_pruning, device)

lib/rank_utils.py:155

Functionrank_reduction_weight_wrapper

(args, model, tokenizer, rank_pruning, device)

lib/rank_utils.py:179

Functionrank_reduction_weight_wrapper_selective

(args, model, tokenizer, rank_pruning, device)

lib/rank_utils.py:207

Functionrank_reduction_weight_wrapper_selective_eval

(args, model, tokenizer, rank_pruning, device)

lib/rank_utils.py:246

welore_downstream_finetune.py:56

Functionsave_dict

(item, filename)

Functionseed_worker

lib/downstream_utils.py:258

Methodset_decoder

(self, decoder)

peft_pretraining/modeling_llama.py:625

Methodset_input_embeddings

peft_pretraining/modeling_llama.py:447

Methodset_input_embeddings

peft_pretraining/modeling_llama.py:616

Methodset_input_embeddings

peft_pretraining/modeling_llama.py:790

Methodset_output_embeddings

(self, new_embeddings)

peft_pretraining/modeling_llama.py:622

Functionset_seed

lib/data_utils.py:9

FunctionshuffleDict

Performs a single optimization step. Arguments: closure (callable, optional): A closure that reevaluates the model

galore_torch/q_galore_adamw8bit.py:58

Performs a single optimization step Arguments: closure (callable, optional): A closure that reevaluates the model

galore_torch/adafactor.py:161

Performs a single optimization step. Arguments: closure (callable, optional): A closure that reevaluates the model

galore_torch/adamw8bit.py:13

Functionuniform_rank_pruning

(args, pruning_ratio, layers_singular_value, logger)

Functionuniform_rank_pruning_exp2

(args, pruning_ratio, layers_singular_value, file_name)

Functionweight_thresold_rank_pruning

Given a rank thresold, normalize the singular values and prune each layer under the rank_thresold

← previous101–139 of 139, ranked by callers