{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Relevancy Analysis" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "
\n", "\n", "This tutorial is available as an IPython notebook at [Malaya/example/relevancy](https://github.com/huseinzol05/Malaya/tree/master/example/relevancy).\n", " \n", "
" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "
\n", "\n", "This module only trained on standard language structure, so it is not save to use it for local language structure.\n", " \n", "
" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "import logging\n", "\n", "logging.basicConfig(level=logging.INFO)" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "INFO:numexpr.utils:NumExpr defaulting to 8 threads.\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "CPU times: user 5.8 s, sys: 1.1 s, total: 6.9 s\n", "Wall time: 7.91 s\n" ] } ], "source": [ "%%time\n", "import malaya" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### labels supported\n", "\n", "Default labels for relevancy module." ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "['not relevant', 'relevant']" ] }, "execution_count": 2, "metadata": {}, "output_type": "execute_result" } ], "source": [ "malaya.relevancy.label" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Explanation\n", "\n", "Positive relevancy: The article or piece of text is relevant, tendency is high to become not a fake news. Can be a positive or negative sentiment.\n", "\n", "Negative relevancy: The article or piece of text is not relevant, tendency is high to become a fake news. Can be a positive or negative sentiment.\n", "\n", "**Right now relevancy module only support deep learning model**." ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [], "source": [ "negative_text = 'Roti Massimo Mengandungi DNA Babi. Roti produk Massimo keluaran Syarikat The Italian Baker mengandungi DNA babi. Para pengguna dinasihatkan supaya tidak memakan produk massimo. Terdapat pelbagai produk roti keluaran syarikat lain yang boleh dimakan dan halal. Mari kita sebarkan berita ini supaya semua rakyat Malaysia sedar dengan apa yang mereka makna setiap hari. Roti tidak halal ada DNA babi jangan makan ok.'\n", "positive_text = 'Jabatan Kemajuan Islam Malaysia memperjelaskan dakwaan sebuah mesej yang dikitar semula, yang mendakwa kononnya kod E dikaitkan dengan kandungan lemak babi sepertimana yang tular di media sosial. . Tular: November 2017 . Tular: Mei 2014 JAKIM ingin memaklumkan kepada masyarakat berhubung maklumat yang telah disebarkan secara meluas khasnya melalui media sosial berhubung kod E yang dikaitkan mempunyai lemak babi. Untuk makluman, KOD E ialah kod untuk bahan tambah (aditif) dan ianya selalu digunakan pada label makanan di negara Kesatuan Eropah. Menurut JAKIM, tidak semua nombor E yang digunakan untuk membuat sesuatu produk makanan berasaskan dari sumber yang haram. Sehubungan itu, sekiranya sesuatu produk merupakan produk tempatan dan mendapat sijil Pengesahan Halal Malaysia, maka ia boleh digunakan tanpa was-was sekalipun mempunyai kod E-kod. Tetapi sekiranya produk tersebut bukan produk tempatan serta tidak mendapat sijil pengesahan halal Malaysia walaupun menggunakan e-kod yang sama, pengguna dinasihatkan agar berhati-hati dalam memilih produk tersebut.'" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### List available Transformer models" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "INFO:malaya.relevancy:trained on 90% dataset, tested on another 10% test set, dataset at https://github.com/huseinzol05/malaya/blob/master/session/relevancy/download-data.ipynb\n" ] }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
Size (MB)Quantized Size (MB)macro precisionmacro recallmacro f1-scoremax length
bert425.6111.000.893200.891950.89256512.0
tiny-bert57.415.400.871790.863240.86695512.0
albert48.612.800.897980.860080.87209512.0
tiny-albert22.45.980.821570.834100.82416512.0
xlnet446.6118.000.927070.921030.92381512.0
alxlnet46.813.300.911350.904460.90758512.0
bigbird458.0116.000.880930.868320.873521024.0
tiny-bigbird65.016.900.865580.858710.861761024.0
fastformer458.0116.000.923870.910640.916162048.0
tiny-fastformer77.319.700.856550.863370.859252048.0
\n", "
" ], "text/plain": [ " Size (MB) Quantized Size (MB) macro precision \\\n", "bert 425.6 111.00 0.89320 \n", "tiny-bert 57.4 15.40 0.87179 \n", "albert 48.6 12.80 0.89798 \n", "tiny-albert 22.4 5.98 0.82157 \n", "xlnet 446.6 118.00 0.92707 \n", "alxlnet 46.8 13.30 0.91135 \n", "bigbird 458.0 116.00 0.88093 \n", "tiny-bigbird 65.0 16.90 0.86558 \n", "fastformer 458.0 116.00 0.92387 \n", "tiny-fastformer 77.3 19.70 0.85655 \n", "\n", " macro recall macro f1-score max length \n", "bert 0.89195 0.89256 512.0 \n", "tiny-bert 0.86324 0.86695 512.0 \n", "albert 0.86008 0.87209 512.0 \n", "tiny-albert 0.83410 0.82416 512.0 \n", "xlnet 0.92103 0.92381 512.0 \n", "alxlnet 0.90446 0.90758 512.0 \n", "bigbird 0.86832 0.87352 1024.0 \n", "tiny-bigbird 0.85871 0.86176 1024.0 \n", "fastformer 0.91064 0.91616 2048.0 \n", "tiny-fastformer 0.86337 0.85925 2048.0 " ] }, "execution_count": 3, "metadata": {}, "output_type": "execute_result" } ], "source": [ "malaya.relevancy.available_transformer()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Load Transformer model\n", "\n", "```python\n", "def transformer(model: str = 'xlnet', quantized: bool = False, **kwargs):\n", " \"\"\"\n", " Load Transformer relevancy model.\n", "\n", " Parameters\n", " ----------\n", " model : str, optional (default='bert')\n", " Model architecture supported. Allowed values:\n", "\n", " * ``'bert'`` - Google BERT BASE parameters.\n", " * ``'tiny-bert'`` - Google BERT TINY parameters.\n", " * ``'albert'`` - Google ALBERT BASE parameters.\n", " * ``'tiny-albert'`` - Google ALBERT TINY parameters.\n", " * ``'xlnet'`` - Google XLNET BASE parameters.\n", " * ``'alxlnet'`` - Malaya ALXLNET BASE parameters.\n", " * ``'bigbird'`` - Google BigBird BASE parameters.\n", " * ``'tiny-bigbird'`` - Malaya BigBird BASE parameters.\n", " * ``'fastformer'`` - FastFormer BASE parameters.\n", " * ``'tiny-fastformer'`` - FastFormer TINY parameters.\n", "\n", " quantized : bool, optional (default=False)\n", " if True, will load 8-bit quantized model.\n", " Quantized model not necessary faster, totally depends on the machine.\n", "\n", " Returns\n", " -------\n", " result: model\n", " List of model classes:\n", "\n", " * if `bert` in model, will return `malaya.model.bert.MulticlassBERT`.\n", " * if `xlnet` in model, will return `malaya.model.xlnet.MulticlassXLNET`.\n", " * if `bigbird` in model, will return `malaya.model.xlnet.MulticlassBigBird`.\n", " * if `fastformer` in model, will return `malaya.model.fastformer.MulticlassFastFormer`.\n", " \"\"\"\n", "```" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "WARNING:tensorflow:From /Users/huseinzolkepli/Documents/Malaya/malaya/function/__init__.py:112: The name tf.gfile.GFile is deprecated. Please use tf.io.gfile.GFile instead.\n", "\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "WARNING:tensorflow:From /Users/huseinzolkepli/Documents/Malaya/malaya/function/__init__.py:112: The name tf.gfile.GFile is deprecated. Please use tf.io.gfile.GFile instead.\n", "\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "WARNING:tensorflow:From /Users/huseinzolkepli/Documents/Malaya/malaya/function/__init__.py:114: The name tf.GraphDef is deprecated. Please use tf.compat.v1.GraphDef instead.\n", "\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "WARNING:tensorflow:From /Users/huseinzolkepli/Documents/Malaya/malaya/function/__init__.py:114: The name tf.GraphDef is deprecated. Please use tf.compat.v1.GraphDef instead.\n", "\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "WARNING:tensorflow:From /Users/huseinzolkepli/Documents/Malaya/malaya/function/__init__.py:107: The name tf.InteractiveSession is deprecated. Please use tf.compat.v1.InteractiveSession instead.\n", "\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "WARNING:tensorflow:From /Users/huseinzolkepli/Documents/Malaya/malaya/function/__init__.py:107: The name tf.InteractiveSession is deprecated. Please use tf.compat.v1.InteractiveSession instead.\n", "\n" ] } ], "source": [ "model = malaya.relevancy.transformer(model = 'tiny-bigbird')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Load Quantized model\n", "\n", "To load 8-bit quantized model, simply pass `quantized = True`, default is `False`.\n", "\n", "We can expect slightly accuracy drop from quantized model, and not necessary faster than normal 32-bit float model, totally depends on machine." ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [], "source": [ "quantized_model = malaya.relevancy.transformer(model = 'alxlnet', quantized = True)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Predict batch of strings\n", "\n", "```python\n", "def predict(self, strings: List[str]):\n", " \"\"\"\n", " classify list of strings.\n", "\n", " Parameters\n", " ----------\n", " strings: List[str]\n", "\n", " Returns\n", " -------\n", " result: List[str]\n", " \"\"\"\n", "```" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "CPU times: user 2.04 s, sys: 520 ms, total: 2.56 s\n", "Wall time: 1.23 s\n" ] }, { "data": { "text/plain": [ "['not relevant', 'relevant']" ] }, "execution_count": 7, "metadata": {}, "output_type": "execute_result" } ], "source": [ "%%time\n", "\n", "model.predict([negative_text, positive_text])" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "CPU times: user 5.08 s, sys: 823 ms, total: 5.91 s\n", "Wall time: 2.96 s\n" ] }, { "data": { "text/plain": [ "['not relevant', 'relevant']" ] }, "execution_count": 8, "metadata": {}, "output_type": "execute_result" } ], "source": [ "%%time\n", "\n", "quantized_model.predict([negative_text, positive_text])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Predict batch of strings with probability\n", "\n", "```python\n", "def predict_proba(self, strings: List[str]):\n", " \"\"\"\n", " classify list of strings and return probability.\n", "\n", " Parameters\n", " ----------\n", " strings : List[str]\n", "\n", " Returns\n", " -------\n", " result: List[dict[str, float]]\n", " \"\"\"\n", "```" ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "CPU times: user 1.46 s, sys: 403 ms, total: 1.86 s\n", "Wall time: 319 ms\n" ] }, { "data": { "text/plain": [ "[{'not relevant': 0.9896912, 'relevant': 0.010308762},\n", " {'not relevant': 0.007830339, 'relevant': 0.9921697}]" ] }, "execution_count": 9, "metadata": {}, "output_type": "execute_result" } ], "source": [ "%%time\n", "\n", "model.predict_proba([negative_text, positive_text])" ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "CPU times: user 2.98 s, sys: 386 ms, total: 3.37 s\n", "Wall time: 583 ms\n" ] }, { "data": { "text/plain": [ "[{'not relevant': 0.9999988, 'relevant': 1.2511766e-06},\n", " {'not relevant': 9.157779e-06, 'relevant': 0.9999908}]" ] }, "execution_count": 10, "metadata": {}, "output_type": "execute_result" } ], "source": [ "%%time\n", "\n", "quantized_model.predict_proba([negative_text, positive_text])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Open relevancy visualization dashboard\n", "\n", "Default when you call `predict_words` it will open a browser with visualization dashboard, you can disable by `visualization=False`.\n", "\n", "```python\n", "def predict_words(\n", " self,\n", " string: str,\n", " method: str = 'last',\n", " bins_size: float = 0.05,\n", " visualization: bool = True,\n", "):\n", " \"\"\"\n", " classify words.\n", "\n", " Parameters\n", " ----------\n", " string : str\n", " method : str, optional (default='last')\n", " Attention layer supported. Allowed values:\n", "\n", " * ``'last'`` - attention from last layer.\n", " * ``'first'`` - attention from first layer.\n", " * ``'mean'`` - average attentions from all layers.\n", " bins_size: float, optional (default=0.05)\n", " default bins size for word distribution histogram.\n", " visualization: bool, optional (default=True)\n", " If True, it will open the visualization dashboard.\n", "\n", " Returns\n", " -------\n", " dictionary: results\n", " \"\"\"\n", "```\n", "\n", "**This method not available for BigBird models**." ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "quantized_model.predict_words(negative_text)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Vectorize\n", "\n", "Let say you want to visualize sentence / word level in lower dimension, you can use `model.vectorize`,\n", "\n", "```python\n", "def vectorize(self, strings: List[str], method: str = 'first'):\n", " \"\"\"\n", " vectorize list of strings.\n", "\n", " Parameters\n", " ----------\n", " strings: List[str]\n", " method : str, optional (default='first')\n", " Vectorization layer supported. Allowed values:\n", "\n", " * ``'last'`` - vector from last sequence.\n", " * ``'first'`` - vector from first sequence.\n", " * ``'mean'`` - average vectors from all sequences.\n", " * ``'word'`` - average vectors based on tokens.\n", "\n", " Returns\n", " -------\n", " result: np.array\n", " \"\"\"\n", "```" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Sentence level" ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [], "source": [ "texts = [negative_text, positive_text]\n", "r = model.vectorize(texts, method = 'first')" ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "(2, 2)" ] }, "execution_count": 11, "metadata": {}, "output_type": "execute_result" } ], "source": [ "from sklearn.manifold import TSNE\n", "import matplotlib.pyplot as plt\n", "\n", "tsne = TSNE().fit_transform(r)\n", "tsne.shape" ] }, { "cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "plt.figure(figsize = (7, 7))\n", "plt.scatter(tsne[:, 0], tsne[:, 1])\n", "labels = texts\n", "for label, x, y in zip(\n", " labels, tsne[:, 0], tsne[:, 1]\n", "):\n", " label = (\n", " '%s, %.3f' % (label[0], label[1])\n", " if isinstance(label, list)\n", " else label\n", " )\n", " plt.annotate(\n", " label,\n", " xy = (x, y),\n", " xytext = (0, 0),\n", " textcoords = 'offset points',\n", " )" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Word level" ] }, { "cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [], "source": [ "r = quantized_model.vectorize(texts, method = 'word')" ] }, { "cell_type": "code", "execution_count": 14, "metadata": {}, "outputs": [], "source": [ "x, y = [], []\n", "for row in r:\n", " x.extend([i[0] for i in row])\n", " y.extend([i[1] for i in row])" ] }, { "cell_type": "code", "execution_count": 15, "metadata": { "scrolled": true }, "outputs": [ { "data": { "text/plain": [ "(211, 2)" ] }, "execution_count": 15, "metadata": {}, "output_type": "execute_result" } ], "source": [ "tsne = TSNE().fit_transform(y)\n", "tsne.shape" ] }, { "cell_type": "code", "execution_count": 16, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "plt.figure(figsize = (7, 7))\n", "plt.scatter(tsne[:, 0], tsne[:, 1])\n", "labels = x\n", "for label, x, y in zip(\n", " labels, tsne[:, 0], tsne[:, 1]\n", "):\n", " label = (\n", " '%s, %.3f' % (label[0], label[1])\n", " if isinstance(label, list)\n", " else label\n", " )\n", " plt.annotate(\n", " label,\n", " xy = (x, y),\n", " xytext = (0, 0),\n", " textcoords = 'offset points',\n", " )" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Pretty good, the model able to know cluster bottom left as positive relevancy." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Stacking models\n", "\n", "More information, you can read at [https://malaya.readthedocs.io/en/latest/Stack.html](https://malaya.readthedocs.io/en/latest/Stack.html)" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "scrolled": true }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "WARNING:tensorflow:From /Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/albert/tokenization.py:240: The name tf.logging.info is deprecated. Please use tf.compat.v1.logging.info instead.\n", "\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "WARNING:tensorflow:From /Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/albert/tokenization.py:240: The name tf.logging.info is deprecated. Please use tf.compat.v1.logging.info instead.\n", "\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "INFO:tensorflow:loading sentence piece model\n" ] } ], "source": [ "albert = malaya.relevancy.transformer(model = 'albert')" ] }, { "cell_type": "code", "execution_count": 14, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "[{'not relevant': 3.1056952e-06, 'relevant': 0.9999934},\n", " {'not relevant': 0.99982065, 'relevant': 3.868528e-05}]" ] }, "execution_count": 14, "metadata": {}, "output_type": "execute_result" } ], "source": [ "malaya.stack.predict_stack([albert, model], [positive_text, negative_text])" ] } ], "metadata": { "kernelspec": { "display_name": "Python 3", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.7.7" }, "varInspector": { "cols": { "lenName": 16, "lenType": 16, "lenVar": 40 }, "kernels_config": { "python": { "delete_cmd_postfix": "", "delete_cmd_prefix": "del ", "library": "var_list.py", "varRefreshCmd": "print(var_dic_list())" }, "r": { "delete_cmd_postfix": ") ", "delete_cmd_prefix": "rm(", "library": "var_list.r", "varRefreshCmd": "cat(var_dic_list()) " } }, "types_to_exclude": [ "module", "function", "builtin_function_or_method", "instance", "_Feature" ], "window_display": false } }, "nbformat": 4, "nbformat_minor": 2 }