{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "### Zero-Shot Classification in NLP\n", "\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "**Zero-shot text classification is a task in natural language processing where a model is trained on a set of labeled examples but is then able to classify new examples from previously unseen classes.**\n", "\n", "This is similar to the CLIP model in computer vision, but for NLP (text sentences)." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "![zsc](images/zsc.png)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### About the Task\n", "\n", "Zero Shot Classification is the task of predicting a class that wasn't seen by the model during training. This method, which leverages a pre-trained language model, can be thought of as an instance of transfer learning which generally refers to using a model trained for one task in a different application than what it was originally trained for. This is particularly useful for situations where the amount of labeled data is small.\n", "\n", "In zero shot classification, we provide the model with a prompt and a sequence of text that describes what we want our model to do, in natural language. Zero-shot classification excludes any examples of the desired task being completed. This differs from single or few-shot classification, as these tasks include a single or a few examples of the selected task.\n", "\n", "Zero, single and few-shot classification seem to be an emergent feature of large language models. This feature seems to come about around model sizes of +100M parameters. The effectiveness of a model at a zero, single or few-shot task seems to scale with model size, meaning that larger models (models with more trainable parameters or layers) generally do better at this task." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Inference with pipeline\n", "\n" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "No model was supplied, defaulted to facebook/bart-large-mnli and revision c626438 (https://huggingface.co/facebook/bart-large-mnli).\n", "Using a pipeline without specifying a model name and revision in production is not recommended.\n" ] } ], "source": [ "from transformers import pipeline\n", "\n", "pipe = pipeline(\"zero-shot-classification\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Example 1: Customer Service" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [ { "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", "
textcustomer
0I have a problem with my iphone that needs to ...AAAA
1I have a problem with my Mac that is catching ...AA
\n", "
" ], "text/plain": [ " text customer\n", "0 I have a problem with my iphone that needs to ... AAAA\n", "1 I have a problem with my Mac that is catching ... AA" ] }, "execution_count": 2, "metadata": {}, "output_type": "execute_result" } ], "source": [ "import pandas as pd\n", "\n", "data = [{\n", " \"text\" : \"I have a problem with my iphone that needs to be resolved asap!\",\n", " \"customer\" : \"AAAA\"\n", "},\n", "{\n", " \"text\" : \"I have a problem with my Mac that is catching fire right now!\",\n", " \"customer\" : \"AA\"\n", "}]\n", "\n", "df = pd.DataFrame(data)\n", "\n", "df" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "The `multi_class` argument has been deprecated and renamed to `multi_label`. `multi_class` will be removed in a future version of Transformers.\n", "/Users/victorgallego/miniforge3/lib/python3.9/site-packages/torch/utils/data/dataloader.py:643: UserWarning: Length of IterableDataset was reported to be 2 (when accessing len(dataloader)), but 3 samples have been fetched. \n", " warnings.warn(warn_msg)\n", "/Users/victorgallego/miniforge3/lib/python3.9/site-packages/torch/utils/data/dataloader.py:643: UserWarning: Length of IterableDataset was reported to be 2 (when accessing len(dataloader)), but 4 samples have been fetched. \n", " warnings.warn(warn_msg)\n", "/Users/victorgallego/miniforge3/lib/python3.9/site-packages/torch/utils/data/dataloader.py:643: UserWarning: Length of IterableDataset was reported to be 2 (when accessing len(dataloader)), but 5 samples have been fetched. \n", " warnings.warn(warn_msg)\n", "/Users/victorgallego/miniforge3/lib/python3.9/site-packages/torch/utils/data/dataloader.py:643: UserWarning: Length of IterableDataset was reported to be 2 (when accessing len(dataloader)), but 6 samples have been fetched. \n", " warnings.warn(warn_msg)\n", "/Users/victorgallego/miniforge3/lib/python3.9/site-packages/torch/utils/data/dataloader.py:643: UserWarning: Length of IterableDataset was reported to be 2 (when accessing len(dataloader)), but 7 samples have been fetched. \n", " warnings.warn(warn_msg)\n", "/Users/victorgallego/miniforge3/lib/python3.9/site-packages/torch/utils/data/dataloader.py:643: UserWarning: Length of IterableDataset was reported to be 2 (when accessing len(dataloader)), but 8 samples have been fetched. \n", " warnings.warn(warn_msg)\n", "/Users/victorgallego/miniforge3/lib/python3.9/site-packages/torch/utils/data/dataloader.py:643: UserWarning: Length of IterableDataset was reported to be 2 (when accessing len(dataloader)), but 9 samples have been fetched. \n", " warnings.warn(warn_msg)\n", "/Users/victorgallego/miniforge3/lib/python3.9/site-packages/torch/utils/data/dataloader.py:643: UserWarning: Length of IterableDataset was reported to be 2 (when accessing len(dataloader)), but 10 samples have been fetched. \n", " warnings.warn(warn_msg)\n", "/Users/victorgallego/miniforge3/lib/python3.9/site-packages/torch/utils/data/dataloader.py:643: UserWarning: Length of IterableDataset was reported to be 2 (when accessing len(dataloader)), but 11 samples have been fetched. \n", " warnings.warn(warn_msg)\n", "/Users/victorgallego/miniforge3/lib/python3.9/site-packages/torch/utils/data/dataloader.py:643: UserWarning: Length of IterableDataset was reported to be 2 (when accessing len(dataloader)), but 12 samples have been fetched. \n", " warnings.warn(warn_msg)\n" ] }, { "data": { "text/plain": [ "[{'sequence': 'I have a problem with my iphone that needs to be resolved asap!',\n", " 'labels': ['urgent', 'phone', 'computer', 'not urgent', 'tablet', 'Mac'],\n", " 'scores': [0.522032618522644,\n", " 0.4575027823448181,\n", " 0.014244725927710533,\n", " 0.002681280020624399,\n", " 0.0021490638609975576,\n", " 0.0013895597076043487]},\n", " {'sequence': 'I have a problem with my Mac that is catching fire right now!',\n", " 'labels': ['computer', 'Mac', 'urgent', 'not urgent', 'phone', 'tablet'],\n", " 'scores': [0.45493218302726746,\n", " 0.3786928951740265,\n", " 0.1611764281988144,\n", " 0.0029544392600655556,\n", " 0.0011420187074691057,\n", " 0.0011020548408851027]}]" ] }, "execution_count": 4, "metadata": {}, "output_type": "execute_result" } ], "source": [ "pipe(df.text.tolist(),\n", " candidate_labels=[\"urgent\", \"not urgent\", \"phone\", \"tablet\", \"computer\", \"Mac\"],\n", " multi_class=False\n", ")" ] }, { "cell_type": "code", "execution_count": 19, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "The `multi_class` argument has been deprecated and renamed to `multi_label`. `multi_class` will be removed in a future version of Transformers.\n" ] }, { "data": { "text/plain": [ "{'sequence': 'I have a problem with my iphone that needs to be resolved asap!',\n", " 'labels': ['urgent', 'phone', 'computer', 'not urgent', 'tablet'],\n", " 'scores': [0.9987171292304993,\n", " 0.9945850372314453,\n", " 0.18988825380802155,\n", " 0.0007674169610254467,\n", " 0.0003826009633485228]}" ] }, "execution_count": 19, "metadata": {}, "output_type": "execute_result" } ], "source": [ "pipe(\"I have a problem with my iphone that needs to be resolved asap!\",\n", " candidate_labels=[\"urgent\", \"not urgent\", \"phone\", \"tablet\", \"computer\"],\n", ")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Notice all the probabilities (scores) add up to 1.0. This is because the model is trained to output a probability distribution over the classes. The class with the highest probability is \"urgent\"." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Example 2: Activity classification" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "{'sequence': 'one day I will see the world',\n", " 'labels': ['travel', 'dancing', 'cooking'],\n", " 'scores': [0.9938651919364929, 0.0032738191075623035, 0.0028610145673155785]}" ] }, "execution_count": 5, "metadata": {}, "output_type": "execute_result" } ], "source": [ "sequence_to_classify = \"one day I will see the world\"\n", "candidate_labels = ['travel', 'cooking', 'dancing']\n", "pipe(sequence_to_classify, candidate_labels)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Multi-Label Classification" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "If more than one candidate label can be correct, pass `multi_label=True` to calculate each class independently:" ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "{'sequence': 'one day I will see the world',\n", " 'labels': ['travel', 'exploration', 'dancing', 'cooking'],\n", " 'scores': [0.994511067867279,\n", " 0.9383878111839294,\n", " 0.005706228781491518,\n", " 0.0018192909192293882]}" ] }, "execution_count": 6, "metadata": {}, "output_type": "execute_result" } ], "source": [ "candidate_labels = ['travel', 'cooking', 'dancing', 'exploration']\n", "pipe(sequence_to_classify, candidate_labels, multi_label=True)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Notice now the probabilities are independent of each other, and the model is able to predict multiple classes at the same time (i.e., they do not add up to 1.0)." ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "The `multi_class` argument has been deprecated and renamed to `multi_label`. `multi_class` will be removed in a future version of Transformers.\n" ] }, { "data": { "text/plain": [ "{'sequence': 'I have a problem with my iphone that needs to be resolved asap!',\n", " 'labels': ['urgent', 'phone', 'computer', 'not urgent', 'tablet'],\n", " 'scores': [0.9987171292304993,\n", " 0.9945850372314453,\n", " 0.18988825380802155,\n", " 0.0007674169610254467,\n", " 0.0003826009633485228]}" ] }, "execution_count": 5, "metadata": {}, "output_type": "execute_result" } ], "source": [ "pipe(\"I have a problem with my iphone that needs to be resolved asap!\",\n", " candidate_labels=[\"urgent\", \"not urgent\", \"phone\", \"tablet\", \"computer\"],\n", " multi_class=True\n", ")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Available zero-shot classification models\n", "\n", "All the models can be checked at https://huggingface.co/models?pipeline_tag=zero-shot-classification&sort=downloads\n", "\n", "The main differences:\n", "\n", "* Size (in number of parameters): bigger models tend to be more accurate, but also slower and more resource-intensive (remember the trade-off from computer vision models, it happens the same in NLP).\n", "* Languages: most models are trained only in English, but some other are specialized in other languages, or are even multilingual." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Multi-lingual Zero-Shot Classification 🌍" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "If a model has the \"XLM\" as part of its name, it means it is a multi-lingual model. This means it can understand text in multiple languages. This is particularly useful for companies that operate in multiple countries and need to understand and generate text in multiple languages.\n", "\n", "(Apart from \"XLM\", there are many more models that are multilingual, the previous is just an example of a popular family of multilingual models)." ] }, { "cell_type": "markdown", "metadata": {}, "source": [] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [], "source": [ "pipe = pipeline('zero-shot-classification', model=\"vicgalle/xlm-roberta-large-xnli-anli\")" ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "{'sequence': 'Algún día iré a ver el mundo',\n", " 'labels': ['viaje', 'danza', 'cocina'],\n", " 'scores': [0.9991760849952698, 0.0004178229719400406, 0.00040599776548333466]}" ] }, "execution_count": 9, "metadata": {}, "output_type": "execute_result" } ], "source": [ "sequence_to_classify = \"Algún día iré a ver el mundo\"\n", "candidate_labels = ['viaje', 'cocina', 'danza']\n", "\n", "pipe(sequence_to_classify, candidate_labels)" ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [], "source": [ "# Function to plot results\n", "import seaborn as sns\n", "import matplotlib.pyplot as plt\n", "sns.set_style(\"whitegrid\")\n", "sns.set_context(\"notebook\")\n", "def plot_output(output):\n", " scores = output[\"scores\"]\n", " colors = list(map(lambda color: sns.palettes.light_palette(color)[2],[\"red\" if score > .5 else \"blue\" for score in scores]))\n", " plt.figure(dpi=100)\n", " plt.bar(output[\"labels\"], scores, color = colors )\n", " plt.xlabel(\"Labels\")\n", " plt.ylabel(\"Probability\")\n", " plt.title(output[\"sequence\"])\n", " n = list(range(len(scores)))\n", " for i in n:\n", " plt.annotate(f\"{100*scores[i]:.1f}%\", xy=(n[i],scores[i]*.8), ha='center', va='bottom')\n", "\n", " plt.show()" ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "sequence_to_classify = \"La película empezó bien pero terminó siendo un desastre.\"\n", "candidate_labels = \"positivo negativo neutro\".split()\n", "output = pipe (sequence_to_classify, candidate_labels)\n", "\n", "plot_output(output)" ] }, { "cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "sequence_to_classify = \"La película empezó fatal pero terminó siendo una maravilla.\"\n", "candidate_labels = \"positivo negativo neutro\".split()\n", "output = pipe(sequence_to_classify, candidate_labels)\n", "plot_output(output)" ] }, { "cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "sequence_to_classify = \"La película estuvo normal, no me encantó, pero tampoco me aburrió.\"\n", "candidate_labels = \"positivo negativo neutro\".split()\n", "output = pipe(sequence_to_classify, candidate_labels)\n", "\n", "plot_output(output)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Notice we can even write the class labels in another language. Probabilities shall no change significantly." ] }, { "cell_type": "code", "execution_count": 14, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "sequence_to_classify = \"La película empezó fatal pero terminó siendo una maravilla.\"\n", "candidate_labels = \"positive negative neutral\".split()\n", "output = pipe(sequence_to_classify, candidate_labels)\n", "plot_output(output)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Topic classification" ] }, { "cell_type": "code", "execution_count": 15, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "sequence_to_classify = \"La farmacéutica GAVI, para quien no lo sepa, es propiedad de la fundación BILL & MELINDA GATES, los especialistas en vacunas fallidas que tantas víctimas han causado alrededor del mundo. India les ha expulsado y denunciado. África aún acarrea sus consecuencias.\"\n", "\n", "candidate_labels = \"music cinema politics science art\".split()\n", "output = pipe(sequence_to_classify, candidate_labels)\n", "plot_output(output)" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "sequence_to_classify = \"La farmacéutica GAVI, para quien no lo sepa, es propiedad de la fundación BILL & MELINDA GATES, los especialistas en vacunas fallidas que tantas víctimas han causado alrededor del mundo. India les ha expulsado y denunciado. África aún acarrea sus consecuencias.\"\n", "\n", "candidate_labels = \"music cinema politics science art\".split()\n", "output = pipe(sequence_to_classify, candidate_labels)\n", "plot_output(output)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### How to check the number of parameters of a transformers model?\n", "\n", "In its page of the Model Hub, notice the parameter count to the right:\n", "\n", "\n", "![params](images/model_hub.png)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Exercise: Brainstorm an NLP application in which you could use zero-shot classification" ] }, { "cell_type": "code", "execution_count": 24, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "# Detecting threats\n", "\n", "sequence_to_classify = \"We should give him what he deserves.\"\n", "\n", "candidate_labels = [\"An offensive comment\", \"A non-offensive comment\"]\n", "output = pipe(sequence_to_classify, candidate_labels)\n", "plot_output(output)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Extra: how does zero-shot classification work under the hood?" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "**The NLP models to do zero-shot classification are pretrained on a generic task that is called \"Natural Language Inference\".**\n", "\n", "In this task, we consider two sentences as input:\n", "* A premise.\n", "* An hypothesis.\n", "\n", "The task is to classify whether the hypothesis is TRUE (entailment), FALSE (contradiction), or NEUTRAL, given the premise. That is, is a 3-label classification problem. See the table for a few examples:\n", "\n", "![examples](https://joeddav.github.io/blog/images/zsl/nli-examples.png)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Inside the model, we feed as inputs both the premise and the hypothesis, and then the final NN layer just classifies into contradiction, entailment, or neutral." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "##### How to turn this problem into zero-shot classification?\n", "\n", "Just take the sentence we want to classify as the premise, and use one different hypothesis for each label we want to classify. Then, the model will output the probabilities of each label being true, given the sentence.\n", "\n", "For example:\n", "\n", "* Premise: \"I hate this movie, it is so boring.\"\n", "\n", "Let's say we want to zero-shot classify into positive or negative sentiment. We can use the following hypotheses:\n", "\n", "* Hypothesis 1: \"This is positive.\"\n", "* Hypothesis 2: \"This is negative.\"\n", "\n", "Then, the model will output the probabilities of each hypothesis being true (entailment), given the sentence. This is how zero-shot classification works under the hood." ] } ], "metadata": { "kernelspec": { "display_name": "base", "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": "3.9.7" } }, "nbformat": 4, "nbformat_minor": 2 }