how to run BERT

[maggiezha]

Could you please give some instructions on how to run lesson29-BERT? I tried to do "pip3 install keras-bert" which was from the link to the other work that you are based on, then run "python3 tokenizer.py" but got: from .bert import TOKEN_CLS, TOKEN_SEP, TOKEN_UNK ModuleNotFoundError: No module named '__ main__.bert'; '__ main__' is not a package or if I tried "python3 bert.py" but got: from .layers import (get_inputs, get_embedding, TokenEmbedding, EmbeddingSimilarity, Masked, Extract) ModuleNotFoundError: No module named '__ main__.layers'; '__ main__' is not a package

[dragen1860]

@ash3n

[ash3n]

The demo code from keras-bert should still work if you rename a few dependencies.

Delete the period in "from .layers import" in bert.py and try running this:

from tensorflow import keras
from bert import get_base_dict, get_model, gen_batch_inputs


# A toy input example
sentence_pairs = [
    [['all', 'work', 'and', 'no', 'play'], ['makes', 'jack', 'a', 'dull', 'boy']],
    [['from', 'the', 'day', 'forth'], ['my', 'arm', 'changed']],
    [['and', 'a', 'voice', 'echoed'], ['power', 'give', 'me', 'more', 'power']],
]


# Build token dictionary
token_dict = get_base_dict()  # A dict that contains some special tokens
for pairs in sentence_pairs:
    for token in pairs[0] + pairs[1]:
        if token not in token_dict:
            token_dict[token] = len(token_dict)
token_list = list(token_dict.keys())  # Used for selecting a random word


# Build & train the model
model = get_model(
    token_num=len(token_dict),
    head_num=5,
    transformer_num=12,
    embed_dim=25,
    feed_forward_dim=100,
    seq_len=20,
    pos_num=20,
    dropout_rate=0.05,
)
model.summary()

def _generator():
    while True:
        yield gen_batch_inputs(
            sentence_pairs,
            token_dict,
            token_list,
            seq_len=20,
            mask_rate=0.3,
            swap_sentence_rate=1.0,
        )

model.fit_generator(
    generator=_generator(),
    steps_per_epoch=1000,
    epochs=100,
    validation_data=_generator(),
    validation_steps=100,
    callbacks=[
        keras.callbacks.EarlyStopping(monitor='val_loss', patience=5)
    ],
)


# Use the trained model
inputs, output_layer = get_model(
    token_num=len(token_dict),
    head_num=5,
    transformer_num=12,
    embed_dim=25,
    feed_forward_dim=100,
    seq_len=20,
    pos_num=20,
    dropout_rate=0.05,
    training=False,      # The input layers and output layer will be returned if `training` is `False`
    trainable=False,     # Whether the model is trainable. The default value is the same with `training`
    output_layer_num=4,  # The number of layers whose outputs will be concatenated as a single output.
                         # Only available when `training` is `False`.
)

I'm pushing a main.py for lesson29-BERT soon so hopefully that'll make it more convenient to use.