Functions — github.com/ImKeTT/AdaVAE

hub / github.com/ImKeTT/AdaVAE / functions

Functions280 in github.com/ImKeTT/AdaVAE

⨍ Functions 280◇ Types & classes 57

↓ 214 callersMethodinfo

(self, text)

src/logger.py:30

↓ 89 callersMethodmean

(self, embs, lens)

dialogue/eval_dialog_response.py:75

↓ 22 callersMethodfrom_file

(file_path: str)

src/data.py:94

↓ 21 callersMethod_read_tsv

Reads a tab separated value file.

low_nlu/utils_glue.py:83

↓ 16 callersMethod__init__

(self, config, AdapterConfig)

src/adapters/vae.py:849

↓ 14 callersFunctionnum_params

(model)

src/utils.py:159

↓ 12 callersMethodeval

(self)

src/adapters/common.py:148

↓ 12 callersFunctiontokenize

(texts, tokenizer, device, args)

src/utils.py:148

↓ 11 callersFunctioninit_para_frompretrained

(m, pm, share_para=False)

src/utils.py:167

↓ 11 callersMethodstep

(self, optimizer, loss)

controlgen/oracle_cls.py:65

↓ 10 callersFunctionunfreeze_GPT2_adapters

(GPT2_model: nn.Module, Adapters: list)

src/utils.py:188

↓ 9 callersMethod__init__

(self, ada_config, config)

src/adapters/common.py:283

↓ 9 callersMethodmean

(self, embs, lens)

dialogue/eval_dialog_multi_response.py:74

↓ 9 callersMethodtrain

(self, mode: bool = True)

src/adapters/common.py:137

↓ 8 callersMethodreparameterize

(self, mean, logvar, z=None, ns=0)

src/adapters/vae.py:1517

↓ 8 callersFunctionsimple_accuracy

(preds, labels)

low_nlu/utils_glue.py:613

↓ 7 callersFunctionanalogy

(args, ada_config, model, tokenizer, device, batch_triple, f, top_k=100, top_p=0.5)

src/test.py:229

↓ 7 callersFunctioncompute_loss

:param device: :param model: :param input_tokens: input word ids :param mask: input mask :param x_tokens: target sequence :p

src/adaVAE.py:143

↓ 7 callersFunctioninterpolate

(args, ada_config, model, tokenizer, device, batch_pair, f, num_interpolation_steps=10, top_k=100, top_p=0.5)

src/test.py:195

↓ 7 callersMethodreparameterize

(self, mean, logvar, z=None)

src/adapters/vae.py:1365

↓ 7 callersFunctionsample_sequence

(model, length, z=None, batch_size=None, temperature=1, top_k=100, top_p=0.95, device='cud

src/utils.py:88

↓ 6 callersMethod_shape

(self, x, bsz)

src/adapters/common.py:467

↓ 4 callersFunctionlinear_schedule

(args)

src/utils.py:208

↓ 4 callersMethodsample_sequence_conditional_batch

(self, past, context)

controlgen/ctrl_gen.py:252

↓ 4 callersFunctionswitch_schedule

Apply LR multiplier before iteration "switch"

src/utils.py:197

↓ 4 callersFunctiontokenize

(texts, tokenizer, device, args)

controlgen/oracle_cls.py:43

↓ 4 callersFunctionval_step

(val_loader)

controlgen/oracle_cls.py:151

↓ 4 callersFunctionval_step

(args, model, val_loader, ada_config, tokenizer, device, save_folder)

src/test.py:300

↓ 4 callersFunctionval_step

(val_loader)

src/adaVAE.py:533

↓ 3 callersMethod_create_examples

Creates examples for the training and dev sets.

low_nlu/utils_glue.py:357

↓ 3 callersFunctionlog_Normal_diag

(x, mean, log_var, average=False, dim=None)

src/adapters/common.py:537

↓ 3 callersFunctionlog_sum_exp

Numerically stable implementation of the operation value.exp().sum(dim, keepdim).log()

src/utils.py:38

↓ 3 callersMethodsample_sequence_conditional_batch

(self, representations, context)

controlgen/ctrl_gen.py:634

↓ 3 callersMethodtokenize

(self, texts)

metrics/cls_ft.py:94

↓ 2 callersMethod_create_examples

Creates examples for the training and dev sets.

low_nlu/utils_glue.py:113

↓ 2 callersMethod_create_examples

Creates examples for the training and dev sets.

low_nlu/utils_glue.py:146

↓ 2 callersMethod_create_examples

Creates examples for the training and dev sets.

low_nlu/utils_glue.py:188

↓ 2 callersMethod_create_examples

Creates examples for the training and dev sets.

low_nlu/utils_glue.py:231

↓ 2 callersMethod_create_examples

Creates examples for the training and dev sets.

low_nlu/utils_glue.py:281

↓ 2 callersMethod_create_examples

Creates examples for the training and dev sets.

low_nlu/utils_glue.py:325

↓ 2 callersMethod_create_examples

Creates examples for the training and dev sets.

low_nlu/utils_glue.py:393

↓ 2 callersMethod_create_examples

Creates examples for the training and dev sets.

low_nlu/utils_glue.py:425

↓ 2 callersMethod_create_examples

Creates examples for the training and dev sets.

low_nlu/utils_glue.py:457

↓ 2 callersFunctionacc_and_f1

(preds, labels)

low_nlu/utils_glue.py:617

↓ 2 callersFunctioncalc_bow

(ref)

dialogue/eval_dialog_multi_response.py:288

↓ 2 callersFunctionchr_only

(s)

dialogue/eval_dialog_multi_response.py:195

↓ 2 callersFunctioncompare_date

whether date2 is no earlier than date1 :param date1: :param date2: :return:

src/utils.py:247

↓ 2 callersMethodcompute

(self, batch)

metrics/cls_ft.py:106

↓ 2 callersFunctioncompute_loss

(device, model, inputs_src, inputs_tgt, labels_tgt, src_attention_mask, tgt_attention_mask, beta, kl_rate, fb)

dialogue/run_spacefusion_gen.py:153

↓ 2 callersMethodembedding

(self, seqs)

dialogue/eval_dialog_response.py:42

↓ 2 callersMethodembedding

(self, seqs)

dialogue/eval_dialog_multi_response.py:42

↓ 2 callersFunctionevaluate

(args, model, tokenizer, logging, eval_dataloader, max_val_batches, eval_output_dir, iter, device)

controlgen/run_vae_ctrl_gen.py:191

↓ 2 callersFunctionevaluate

(args, model_sf, tokenizer, eval_dataloader, logging, device, prefix="", subset="test")

dialogue/run_spacefusion_gen.py:245

↓ 2 callersMethodextrema

computes the value of every single dimension in the word vectors which has the greatest difference from zero. :param seq: seq

dialogue/eval_dialog_response.py:54

↓ 2 callersMethodextrema

computes the value of every single dimension in the word vectors which has the greatest difference from zero. :param seq: seq

dialogue/eval_dialog_multi_response.py:53

↓ 2 callersFunctionfrange_cycle_zero_linear

(n_iter, start=0.0, stop=1.0, n_cycle=4, ratio_increase=0.5, ratio_zero=0.25)

src/utils.py:230

↓ 2 callersFunctiongumbel_softmax

Sample from the Gumbel-Softmax distribution and optionally discretize. Args: logits: [..., n_class] unnormalized log-probs

controlgen/ctrl_gen.py:737

↓ 2 callersMethodids2speaker

(self, ids)

dialogue/spacefusion.py:64

↓ 2 callersMethodkl_loss

(self, mean1, logvar1, mean2, logvar2)

src/adapters/vae.py:1531

↓ 2 callersMethodreset_parameters

(self)

src/adapters/common.py:126

↓ 2 callersMethodsim_bleu

:param ref - a list of tokens of the reference :param hyps - a list of tokens of the hypothesis :return maxbleu - recall ble

dialogue/eval_dialog_multi_response.py:81

↓ 2 callersMethodtest_step

(self, batch)

metrics/cls_ft.py:115

↓ 2 callersFunctiontokenize

(context, response, tokenizer, device, args)

dialogue/run_spacefusion_gen.py:138

↓ 2 callersFunctiontokenize

(input_example, tokenizer, device, max_length)

low_nlu/run_glue.py:119

↓ 2 callersFunctionval_step

(val_loader)

controlgen/run_vae_ctrl_gen.py:524

↓ 2 callersFunctionval_step

(val_loader)

metrics/cls_ft.py:200

↓ 2 callersFunctionval_step

(data_loader)

low_nlu/run_glue.py:346

↓ 1 callersMethod__init__

(self, encoder, decoder, tokenizer_encoder, tokenizer_decoder, args)

controlgen/ctrl_gen.py:14

↓ 1 callersMethod_attn

(self, q, k, v, attention_mask=None, head_mask=None, output_attentions=False)

src/adapters/vae.py:133

↓ 1 callersMethod_attn

(self, q, k, v, attention_mask=None, head_mask=None, output_attentions=False)

src/adapters/vae.py:249

↓ 1 callersMethod_attn

unmasked attention layer for encoder, re-define _atten function

src/adapters/vae.py:549

↓ 1 callersFunction_truncate_seq_pair

Truncates a sequence pair in place to the maximum length.

low_nlu/utils_glue.py:596

↓ 1 callersMethodadv_loss

adversarial loss for wasserstrain distance calculation from Educating Text Autoencoders: Latent Representation Guidance via Denoising

src/adapters/vae.py:1535

↓ 1 callersFunctioncal_rec

(args, ada_config, model, tokenizer, device, eval_dataloader, save_dir=None, sample=False, top_k=100, top_p=0.

src/test.py:266

↓ 1 callersFunctioncompute_loss

(device, model, x_tokens, input_tokens, att_mask, cond_labels)

controlgen/run_vae_ctrl_gen.py:125

↓ 1 callersFunctioncompute_loss

(device, model, input_tokens, att_mask, seg_ids, labels)

low_nlu/run_glue.py:142

↓ 1 callersFunctionconditional_generate

(args, model, tokenizer, label, n_samples, eval_output_dir, device)

controlgen/run_vae_ctrl_gen.py:149

↓ 1 callersMethoddist_batch

(self, vec)

dialogue/spacefusion.py:210

↓ 1 callersMethoddist_pair

(self, a, b)

dialogue/spacefusion.py:207

↓ 1 callersMethoddiv_distinct

distinct-1 distinct-2 metrics for diversity measure proposed by Li et al. "A Diversity-Promoting Objective Function for Neural Conver

dialogue/eval_dialog_response.py:129

↓ 1 callersMethoddiv_distinct

distinct-1 distinct-2 metrics for diversity measure proposed by Li et al. "A Diversity-Promoting Objective Function for Neural Conver

dialogue/eval_dialog_multi_response.py:127

↓ 1 callersFunctiondownload_and_extract

(task, data_dir)

low_nlu/download_glu_data.py:42

↓ 1 callersFunctiondownload_diagnostic

(data_dir)

low_nlu/download_glu_data.py:110

↓ 1 callersMethodeval_cond_ll

compute log p(x|z)

src/adapters/vae.py:1591

↓ 1 callersFunctioneval_dialog_response

based on: https://github.com/guxd/DialogWAE/blob/29f206af05bfe5fe28fec4448e208310a7c9258d/sample.py quoted from the DialogWAE paper: https://

dialogue/eval_dialog_response.py:161

↓ 1 callersMethodeval_inference_dist

this function computes log q(z | x) Args: z: tensor different z points that will be evaluated, with

src/adapters/vae.py:1565

↓ 1 callersFunctioneval_multi_ref

based on: https://github.com/guxd/DialogWAE/blob/29f206af05bfe5fe28fec4448e208310a7c9258d/sample.py path: each line is '\t'.join([src, ref,

dialogue/eval_dialog_multi_response.py:161

↓ 1 callersFunctionformat_mrpc

(data_dir, path_to_data)

low_nlu/download_glu_data.py:55

↓ 1 callersFunctiongenerate

(args, model, save_dir, bsz, tokenizer, device, parallel=False, topk=100, top_p=0.95)

src/test.py:116

↓ 1 callersMethodget_dev_examples

See base class.

low_nlu/utils_glue.py:348

↓ 1 callersFunctionget_tasks

(task_names)

low_nlu/download_glu_data.py:120

↓ 1 callersMethodget_train_examples

See base class.

low_nlu/utils_glue.py:343

↓ 1 callersFunctiongumbel_softmax_sample

(logits, temperature)

controlgen/ctrl_gen.py:760

↓ 1 callersMethodinit_data

(self, dl)

src/data.py:45

↓ 1 callersMethodinit_data

(self, dl)

src/data.py:77

↓ 1 callersMethodinit_data

(self, dl, dataset)

src/data.py:106

↓ 1 callersMethodinit_data

(self, dl)

src/data.py:140

↓ 1 callersMethodkl_loss

(self, mean1, logvar1, mean2, logvar2)

src/adapters/vae.py:1371

↓ 1 callersMethodload_word2vec

(self, path_word2vec)

dialogue/eval_dialog_response.py:24

↓ 1 callersMethodload_word2vec

(self, path_word2vec)

dialogue/eval_dialog_multi_response.py:25

next →1–100 of 280, ranked by callers