Mock Tensorflow

This tutorial is available as an IPython notebook at Malaya/example/mock-tensorflow.

Starting with Malaya 5.0.1 and malaya-boilerplate 0.0.24, Tensorflow is no longer necessary to install and if Tensorflow absent, it will be replaced with mock object.

Let say you installed Malaya on a fresh machine or using a virtual environment,

!~/huggingface/bin/pip3 freeze | grep 'tensorflow'

This virtual environment does not have Tensorflow installed.

import malaya

/home/husein/dev/malaya-boilerplate/malaya_boilerplate/frozen_graph.py:46: UserWarning: Cannot import beam_search_ops from Tensorflow 1, ['malaya.jawi_rumi.deep_model', 'malaya.phoneme.deep_model', 'malaya.rumi_jawi.deep_model', 'malaya.stem.deep_model'] for stemmer will not available to use, make sure Tensorflow 1 version >= 1.15
  warnings.warn(
/home/husein/huggingface/lib/python3.8/site-packages/tqdm/auto.py:22: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html
  from .autonotebook import tqdm as notebook_tqdm
bs4 is not installed, `malaya.text.function.remove_html_tags` will use regex

import tensorflow as tf

tf.__version__

<malaya_boilerplate.Mock at 0x7f9b7cdddca0>

As you can see, everything is a Mock object, what happened if you tried to call a model using Tensorflow as backend?

Call Tensorflow model

malaya.sentiment.available_transformer()

	Size (MB)	Quantized Size (MB)	macro precision	macro recall	macro f1-score
bert	425.6	111.00	0.93182	0.93442	0.93307
tiny-bert	57.4	15.40	0.93390	0.93141	0.93262
albert	48.6	12.80	0.91228	0.91929	0.91540
tiny-albert	22.4	5.98	0.91442	0.91646	0.91521
xlnet	446.6	118.00	0.92390	0.92629	0.92444
alxlnet	46.8	13.30	0.91896	0.92589	0.92198

model = malaya.sentiment.transformer(model = 'alxlnet')

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
Cell In[5], line 1
----> 1 model = malaya.sentiment.transformer(model = 'alxlnet')

File ~/huggingface/lib/python3.8/site-packages/herpetologist/__init__.py:100, in check_type.<locals>.check(*args, **kwargs)
     97     for p, v in kwargs.items():
     98         nested_check(v, p)
--> 100 return func(*args, **kwargs)

File ~/dev/malaya/malaya/sentiment.py:112, in transformer(model, quantized, **kwargs)
    108 if model not in _transformer_availability:
    109     raise ValueError(
    110         'model not supported, please check supported models from `malaya.sentiment.available_transformer()`.'
    111     )
--> 112 return classification.transformer(
    113     module='sentiment-v2',
    114     label=label,
    115     model=model,
    116     quantized=quantized,
    117     **kwargs
    118 )

File ~/dev/malaya/malaya/supervised/classification.py:141, in transformer(module, label, model, sigmoid, quantized, **kwargs)
    122 def transformer(
    123     module,
    124     label,
   (...)
    128     **kwargs,
    129 ):
    130     path = check_file(
    131         file=model,
    132         module=module,
   (...)
    139         **kwargs,
    140     )
--> 141     g = load_graph(path['model'], **kwargs)
    143     if sigmoid:
    144         selected_model = SIGMOID_MODEL[model]

File ~/dev/malaya/malaya/function/__init__.py:38, in load_graph(frozen_graph_filename, **kwargs)
     37 def load_graph(frozen_graph_filename, **kwargs):
---> 38     return frozen_graph.load_graph(package, frozen_graph_filename, **kwargs)

File ~/dev/malaya-boilerplate/malaya_boilerplate/frozen_graph.py:241, in load_graph(package, frozen_graph_filename, **kwargs)
    238 precision_mode = kwargs.get('precision_mode', 'FP32').upper()
    239 logger.debug(f'precision_mode: {precision_mode}')
--> 241 device = get_device(**kwargs)
    242 logger.debug(f'device: {device}')
    244 t5_graph = kwargs.get('t5_graph', False)

File ~/dev/malaya-boilerplate/malaya_boilerplate/frozen_graph.py:111, in get_device(**kwargs)
    107 if device_type not in {'XLA_CPU', 'XLA_CPU_JIT', 'CPU', 'GPU', 'XLA_GPU'}:
    108     raise ValueError(
    109         "`device` from `device:{no}` must one of ['XLA_CPU', 'XLA_CPU_JIT', 'CPU', 'GPU', 'XLA_GPU']"
    110     )
--> 111 gpus = available_gpu()
    113 if 'GPU' in device:
    114     if not len(gpus):

File ~/dev/malaya-boilerplate/malaya_boilerplate/utils.py:117, in available_gpu(refresh)
    108 def available_gpu(refresh=False):
    109     """
    110     Get list of GPUs and memory limit from `tensorflow.python.client.device_lib.list_local_devices()`.
    111
   (...)
    114     result : List[str]
    115     """
--> 117     devices = available_device(refresh=refresh)
    118     return [d for d in devices if 'GPU' in d[0] and 'XLA' not in d[0]]

File ~/dev/malaya-boilerplate/malaya_boilerplate/utils.py:96, in available_device(refresh)
     93 if DEVICES is None and not refresh:
     94     from tensorflow.python.client import device_lib
---> 96     DEVICES = device_lib.list_local_devices()
     97     DEVICES = [
     98         (
     99             i.name.replace('/device:', ''),
   (...)
    102         for i in DEVICES
    103     ]
    105 return DEVICES

File ~/dev/malaya-boilerplate/malaya_boilerplate/__init__.py:48, in Mock.__call__(self, *args, **kwargs)
     47 def __call__(self, *args, **kwargs):
---> 48     raise ValueError(f'{self.parent_name} is not installed. Please install it and try again.')

ValueError: tensorflow is not installed. Please install it and try again.

Call PyTorch model

Starting from malaya 4.9, we are going to focus Pytorch for main deep learning backend.

model = malaya.similarity.semantic.huggingface()

model.predict_proba(['tak suka ayam'], ['hate chicken'])

You're using a T5TokenizerFast tokenizer. Please note that with a fast tokenizer, using the `__call__` method is faster than using a method to encode the text followed by a call to the `pad` method to get a padded encoding.

array([0.86549413], dtype=float32)