Functions — github.com/bangoc123/transformer

CodeHub MCPcopy Create free account

hub / github.com/bangoc123/transformer / functions

Functions51 in github.com/bangoc123/transformer

⨍ Functions 51 ◇ Types & classes 19

↓ 4 callersFunctiongenerate_padding_mask

transformer/layers/generate_mask.py:4

↓ 4 callersMethodpreprocess_sentence

(self, w, max_length)

↓ 4 callersMethodsplitting_head

Splitting item to heads Parameters ---------- x: tensor query/key/value

transformer/layers/multi_head_attention.py:66

↓ 3 callersFunctionffn

(d_ff=2048, d_model=512, activation='relu')

transformer/layers/position_wise_feed_forward_network.py:6

↓ 3 callersFunctiongenerate_positional_encoding

transformer/layers/generate_position.py:13

↓ 2 callersMethodbuild_tokenizer

(self, lang_tokenizer, lang)

↓ 2 callersFunctiongenerate_look_ahead_mask

transformer/layers/generate_mask.py:11

↓ 2 callersFunctiongenerate_mask

transformer/layers/generate_mask.py:16

↓ 2 callersMethodscaled_dot_product_attention

Calculate Attention score Parameters ---------- q: tensor query shape

transformer/layers/multi_head_attention.py:27

↓ 2 callersMethodtokenize

(self, lang_tokenizer, lang, max_length)

↓ 1 callersMethodbuild_dataset

(self, inp_path, targ_path, buffer_size, batch_size, max_length, num_examples)

↓ 1 callersMethodcal_acc

(self, real, pred)

↓ 1 callersFunctioncreateAngleRates

transformer/layers/generate_position.py:5

↓ 1 callersMethoddisplay_samples

(self, num_of_pairs, inp_lines, targ_lines)

↓ 1 callersMethodfit

↓ 1 callersMethodload_dataset

(self, inp_path, targ_path, max_length, num_examples)

↓ 1 callersMethodloss_function

(self, real, pred)

↓ 1 callersMethodpredict

(self, encoder_input, decoder_input, is_train, max_length, end_token)

↓ 1 callersMethodtrain_step

(self, inp, tar)

(self, step_num)

transformer/optimizer.py:11

(self, inp_lang, targ_lang, vocab_folder)

(self, model, optimizer, epochs, checkpoint_folder)

(self, d_model, warmup_steps = 4000)

transformer/optimizer.py:4

(self, n, h, inp_vocab_size, targ_vocab_size, d_model, d_ff, activation, dropout_rate=0.1, eps=0.1)

transformer/model.py:24

(self, n, h, vocab_size, d_model, d_ff, activation, dropout_rate=0.1, eps=0.1)

transformer/layers/decoder.py:8

(self, h, d_model, d_ff, activation, dropout_rate=0.1, eps=0.1)

transformer/layers/decoder_layer.py:7

Parameters ---------- d_model: int the dimension of linear projection of q, k and v h

transformer/layers/multi_head_attention.py:6

(self, n, h, vocab_size, d_model, d_ff, activation, dropout_rate=0.1, eps=0.1)

transformer/layers/encoder.py:8

(self, h, d_model, d_ff, activation, dropout_rate=0.1, eps=0.1)

transformer/layers/encoder_layer.py:7

Functioncal_acc

transformer/model.py:13

(self, encoder_in, decoder_in, is_train, encoder_padding_mask, decoder_look_ahead_mask, decoder_padding_mask)

transformer/model.py:31

Parameters ---------- q: tensor target sentence shape: (..., q_length) encoder_out: tensor output of encoder shape: (...,

transformer/layers/decoder.py:17

Parameters ---------- q: tensor query shape: (..., q_length, d_model) x has d_model to create skip connection with output Ret

transformer/layers/decoder_layer.py:20

Calculate Attention score Parameters ---------- q: tensor query shape

transformer/layers/multi_head_attention.py:97

Parameters ---------- x: tensor input sentence shape: (..., q_length) is_train: bool i

transformer/layers/encoder.py:17

Parameters ---------- x: tensor query shape: (..., q_length, d_model) x has d_model to create skip connection with output is_

transformer/layers/encoder_layer.py:17

Functionloss_function

Functionloss_function

(loss_object, real, pred)

transformer/loss.py:4

Methodtest_call

transformer/tests/model_test.py:32

Methodtest_call

transformer/tests/layers/decoder_test.py:36

Methodtest_call

transformer/tests/layers/decoder_layer_test.py:29

Methodtest_call

transformer/tests/layers/multi_head_attention_test.py:52

Methodtest_call

transformer/tests/layers/encoder_test.py:30

Methodtest_call

transformer/tests/layers/encoder_layer_test.py:28

transformer/tests/layers/position_wise_feed_forward_network_test.py:24

Methodtest_generate_look_ahead_mask

transformer/tests/layers/generate_mask_test.py:25

Methodtest_generate_padding_mask

transformer/tests/layers/generate_mask_test.py:14

Methodtest_generate_positional_encoding

transformer/tests/layers/generate_position_test.py:14

Methodtest_preprocess_sentence

transformer/tests/data_test.py:20

Methodtest_scaled_dot_product_attention

transformer/tests/layers/multi_head_attention_test.py:32

Methodtest_splitting_head

transformer/tests/layers/multi_head_attention_test.py:44