MCPcopy Create free account

hub / github.com/DevSinghSachan/emdr2 / functions

Functions707 in github.com/DevSinghSachan/emdr2

↓ 1 callersFunction_parse_args
Parse entire arguments.
megatron/global_vars.py:84
↓ 1 callersFunction_print_args
Print arguments.
megatron/arguments.py:151
↓ 1 callersMethod_run_split_on_punc
Splits punctuation on a piece of text.
megatron/tokenizer/bert_tokenization.py:263
↓ 1 callersMethod_run_strip_accents
Strips accents from a piece of text.
megatron/tokenizer/bert_tokenization.py:252
↓ 1 callersFunction_set_adlr_autoresume
Initialize ADLR autoresume.
megatron/global_vars.py:137
↓ 1 callersFunction_set_tensorboard_writer
Set tensorboard writer.
megatron/global_vars.py:118
↓ 1 callersFunction_set_timers
Initialize timers.
megatron/global_vars.py:155
↓ 1 callersMethod_step_with_closure
(self, closure)
megatron/fp16/fp16.py:476
↓ 1 callersMethod_tokenize_chinese_chars
Adds whitespace around any CJK character.
megatron/tokenizer/bert_tokenization.py:283
↓ 1 callersFunction_train
Train the model.
tasks/openqa/e2eqa/train_e2eqa.py:415
↓ 1 callersFunction_train
Train the model.
tasks/openqa/dense_retriever/train_dense_retriever.py:254
↓ 1 callersFunctionaccuracy_func_provider
(single_dataset_provider, datapath)
tasks/openqa/e2eqa/train_e2eqa.py:327
↓ 1 callersFunctionaccuracy_func_provider
Provide function that calculates accuracies.
tasks/openqa/dense_retriever/train_dense_retriever.py:395
↓ 1 callersMethodadd_additional_special_tokens
(self, tokens_list)
megatron/tokenizer/tokenizer.py:176
↓ 1 callersMethodadd_block_data
(self, row_id, block_embeds, allow_overwrite=False)
megatron/data/emdr2_index.py:56
↓ 1 callersMethodadd_embed_data
Add the embedding of each block to the underlying FAISS index
megatron/data/emdr2_index.py:164
↓ 1 callersMethodadd_embed_data
Add the embedding of each block to the underlying FAISS index
megatron/data/emdr2_index.py:241
↓ 1 callersMethodadd_item
(self, tensor)
megatron/data/indexed_dataset.py:284
↓ 1 callersMethodallreduce_params
(reduce_after=True, no_scale=False, fp32_allreduce=False)
megatron/model/distributed.py:35
↓ 1 callersFunctionattn_mask_postprocess
(attn_mask)
megatron/model/t5_model.py:35
↓ 1 callersMethodbackward
(ctx, output_grads)
megatron/model/fused_softmax.py:64
↓ 1 callersMethodbackward
(self, loss, retain_graph=False)
megatron/fp16/loss_scaler.py:71
↓ 1 callersFunctionbackward_step
Backward step.
megatron/training.py:165
↓ 1 callersFunctionbert_extended_attention_mask
(attention_mask)
megatron/model/bert_model.py:35
↓ 1 callersFunctionbias_gelu
(bias, y)
megatron/model/fused_bias_gelu.py:32
↓ 1 callersFunctionbias_gelu_back
(g, bias, y)
megatron/model/fused_bias_gelu.py:40
↓ 1 callersFunctionbuild_evidence_index
()
tasks/openqa/dense_retriever/train_dense_retriever.py:378
↓ 1 callersFunctionbuild_sample
Convert to numpy and return a sample consumed by the batch producer.
megatron/data/orqa_wiki_dataset.py:123
↓ 1 callersFunctionbuild_sample
(query_uid, token_ids, token_types, num_tokens, dec_in_ids, dec_out_ids, loss_mask, reference)
tasks/openqa/e2eqa/train_data_utils.py:84
↓ 1 callersFunctionbuild_sample
Convert to numpy and return a sample consumed by the batch producer.
tasks/openqa/dense_retriever/train_data_utils.py:94
↓ 1 callersFunctionbuild_sample
Convert to numpy and return a sample consumed by the batch producer.
tasks/openqa/dense_retriever/evaluation/data.py:142
↓ 1 callersFunctionbuild_tokens_types_paddings_from_ids
Build token types and paddings, trim if needed, and pad if needed.
megatron/data/orqa_wiki_dataset.py:86
↓ 1 callersFunctionbuild_tokens_types_paddings_from_ids
(src_ids, answer_text_ids, max_seq_length, decoder_seq_length,
tasks/openqa/e2eqa/train_data_utils.py:27
↓ 1 callersFunctionbuild_tokens_types_paddings_from_ids
Build token types and paddings, trim if needed, and pad if needed.
tasks/openqa/dense_retriever/evaluation/data.py:107
↓ 1 callersFunctionbuild_tokens_types_paddings_from_text
Build token types and paddings, trim if needed, and pad if needed.
megatron/data/orqa_wiki_dataset.py:68
↓ 1 callersFunctionbuild_tokens_types_paddings_from_text
Build token types and paddings, trim if needed, and pad if needed.
tasks/openqa/e2eqa/train_data_utils.py:10
↓ 1 callersFunctionbuild_tokens_types_paddings_from_text
Build token types and paddings, trim if needed, and pad if needed.
tasks/openqa/dense_retriever/train_data_utils.py:29
↓ 1 callersFunctionbuild_tokens_types_paddings_from_text
Build token types and paddings, trim if needed, and pad if needed.
tasks/openqa/dense_retriever/evaluation/data.py:95
↓ 1 callersFunctioncalculate_matches
Evaluates answers presence in the set of documents. This function is supposed to be used with a large collection of documents and results. It
tasks/openqa/dense_retriever/evaluation/qa_validation.py:29
↓ 1 callersFunctioncheck_checkpoint_args
Ensure fixed arguments for a model are the same for the input arguments and the one retreived frm checkpoint.
megatron/checkpointing.py:44
↓ 1 callersMethodclip_master_grads
Clips fp32 master gradients via ``torch.nn.utils.clip_grad_norm``. Args: max_norm (float or int): max norm of the gradie
megatron/fp16/fp16.py:332
↓ 1 callersFunctionconvert_module
Converts a module's immediate parameters and buffers to dtype.
megatron/fp16/fp16util.py:64
↓ 1 callersFunctionconvert_network
Converts a network's parameters and buffers to dtype.
megatron/fp16/fp16util.py:80
↓ 1 callersMethodconvert_tokens_to_ids
(self, tokens)
megatron/tokenizer/bert_tokenization.py:178
↓ 1 callersMethodconvert_tokens_to_string
Converts a sequence of tokens (string) in a single string.
megatron/tokenizer/bert_tokenization.py:185
↓ 1 callersFunctioncopy_to_model_parallel_region
(input_)
megatron/mpu/mappings.py:144
↓ 1 callersMethoddata_iterator
iterates through data and handles wrap around
megatron/data/samplers.py:132
↓ 1 callersFunctiondense_retriever
(dataset_cls)
tasks/openqa/dense_retriever/run.py:9
↓ 1 callersMethoddtype
(self)
megatron/data/indexed_dataset.py:427
↓ 1 callersMethodencode
(self, csv_line)
tools/create_evidence_indexed_dataset.py:34
↓ 1 callersMethodend_document
(self)
megatron/data/indexed_dataset.py:291
↓ 1 callersMethodentities
Returns a list of named-entity-recognition tags of each token. Returns None if this annotation was not included.
tasks/openqa/dense_retriever/evaluation/tokenizers.py:80
↓ 1 callersFunctionevaluate
Evaluation.
megatron/training.py:324
↓ 1 callersMethodfaiss_wrapper
(self)
tasks/openqa/dense_retriever/evaluation/evaluate.py:42
↓ 1 callersMethodfinalize
(self, index_file)
megatron/data/indexed_dataset.py:314
↓ 1 callersFunctionfinish_beam
(outs, total_score, batchsize, eos_id)
megatron/model/search_strategy.py:103
↓ 1 callersFunctionfinish_mpu_init
()
megatron/initialize.py:41
↓ 1 callersFunctionflatten
(ids, types)
megatron/model/emdr2_model.py:22
↓ 1 callersMethodforward
(ctx, inputs, mask, scale)
megatron/model/fused_softmax.py:54
↓ 1 callersFunctionfp16_to_fp32
Convert fp16 `val` to fp32
megatron/fp16/fp16.py:56
↓ 1 callersFunctionfp32_to_fp16
Convert fp32 `val` to fp16
megatron/fp16/fp16.py:44
↓ 1 callersFunctiongather_from_model_parallel_region
(input_)
megatron/mpu/mappings.py:156
↓ 1 callersFunctiongather_split_1d_tensor
Opposite of above function, gather values from model parallel ranks.
megatron/mpu/random.py:113
↓ 1 callersFunctiongelu_impl
OpenAI's gelu implementation.
megatron/model/utils.py:52
↓ 1 callersMethodgenerate_output
(self, model, query_uid, query_ids_bert, query_types, query_mask_bert, query_ids_t5, query_ids_t5_len)
megatron/model/search_strategy.py:133
↓ 1 callersMethodgenerate_query_vectors
(self, qa_file, split)
tasks/openqa/dense_retriever/evaluation/evaluate.py:56
↓ 1 callersFunctiongenerate_text
()
tasks/openqa/e2eqa/train_e2eqa.py:229
↓ 1 callersFunctionget_args
()
tools/inverted_title_index.py:67
↓ 1 callersFunctionget_args
()
tools/save_emdr2_models.py:44
↓ 1 callersFunctionget_args
()
tools/create_evidence_indexed_dataset.py:53
↓ 1 callersFunctionget_args
()
tools/clean_checkpoints.py:32
↓ 1 callersFunctionget_bias_dropout_add
(training)
megatron/model/transformer.py:404
↓ 1 callersFunctionget_data_parallel_rank
Return my rank for the data parallel group.
megatron/mpu/initialize.py:228
↓ 1 callersMethodget_evidence_dataset
(self)
tasks/openqa/dense_retriever/evaluation/evaluate.py:39
↓ 1 callersMethodget_evidence_embedding
(self, path)
megatron/model/emdr2_model.py:410
↓ 1 callersMethodget_evidence_embedding
(self)
tasks/openqa/dense_retriever/evaluation/evaluate.py:35
↓ 1 callersFunctionget_group_world_size_rank
()
tasks/openqa/dense_retriever/train_dense_retriever.py:23
↓ 1 callersFunctionget_index_group
()
megatron/mpu/initialize.py:294
↓ 1 callersFunctionget_kl_div_retriever
(lm_logits, topk_log_probs, labels, loss_mask)
tasks/openqa/e2eqa/train_e2eqa.py:184
↓ 1 callersFunctionget_learning_rate_scheduler
Build the learning rate scheduler.
megatron/training.py:104
↓ 1 callersFunctionget_loss_and_retriever_utility
This function computes loss in a stable manner, and also computes retriever utility
tasks/openqa/e2eqa/train_e2eqa.py:72
↓ 1 callersMethodget_lr
Learning rate decay functions from: https://openreview.net/pdf?id=BJYwwY9ll pg. 4
megatron/learning_rates.py:51
↓ 1 callersFunctionget_one_epoch_dataloader
(dataset, batch_size=None)
megatron/indexer_emdr2.py:15
↓ 1 callersFunctionget_one_epoch_qa_dataloader
Data loader. Note that batch-size is the local (per GPU) batch-size. NOTE: This dataloader is not distributed !!!
tasks/openqa/dense_retriever/evaluation/data.py:60
↓ 1 callersFunctionget_open_retrieval_batch
(data_iterator)
megatron/data/orqa_wiki_dataset.py:43
↓ 1 callersFunctionget_optimizer
Set up the optimizer.
megatron/training.py:73
↓ 1 callersFunctionget_params_for_weight_decay_optimization
Divide params into with-weight-decay and without-weight-decay groups. Layernorms and baises will have no weight decay but the rest will.
megatron/model/utils.py:64
↓ 1 callersFunctionget_qa_dataset
(qa_file, split)
tasks/openqa/dense_retriever/evaluation/data.py:13
↓ 1 callersFunctionget_rank
()
tasks/openqa/dense_retriever/train_dense_retriever.py:470
↓ 1 callersFunctionget_retrieval_score
()
tasks/openqa/dense_retriever/train_dense_retriever.py:384
↓ 1 callersMethodget_tokentype_embedding
(self)
megatron/model/language_model.py:360
↓ 1 callersMethodget_topk
(self, query_tensor)
megatron/model/emdr2_model.py:435
↓ 1 callersFunctionget_train_group
()
megatron/mpu/initialize.py:288
↓ 1 callersFunctionhas_answer
Check if a document contains an answer string. If `match_type` is string, token matching is done between the text and answer. If `match_type`
tasks/openqa/dense_retriever/evaluation/qa_validation.py:98
↓ 1 callersMethodhas_overflow
(self, params)
megatron/fp16/loss_scaler.py:49
↓ 1 callersMethodhas_overflow_serial
(self, params)
megatron/fp16/loss_scaler.py:120
↓ 1 callersFunctioninfer_dataset_impl
(path)
megatron/data/indexed_dataset.py:35
↓ 1 callersFunctioninit_emdr2_groups
(max_training_rank, world_size)
megatron/mpu/initialize.py:255
↓ 1 callersMethodinit_state_dict_from_bert
Initialize the state from a pretrained BERT model on iteration zero of ICT pretraining
megatron/model/dualencoder_model.py:111
↓ 1 callersMethodinit_state_dict_from_dpr_and_t5
Initialize the state from pre-trained DPR model and pre-trained T5 mode on iteration zero of pretraining
megatron/model/emdr2_model.py:233
← previousnext →201–300 of 707, ranked by callers