Skip to content

Instantly share code, notes, and snippets.

@brockmanmatt

brockmanmatt/wic_self-stuffing_examplesizecomparison.ipynb Secret

Created August 5, 2020 15:46
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save brockmanmatt/0008cf6ae0fea4cb104111472012b864 to your computer and use it in GitHub Desktop.
Save brockmanmatt/0008cf6ae0fea4cb104111472012b864 to your computer and use it in GitHub Desktop.
Download ZIP
WiC_self-stuffing_ExampleSizeComparison.ipynb
Display the source blob
Display the rendered blob
Raw
wic_self-stuffing_examplesizecomparison.ipynb
{
"nbformat": 4,
"nbformat_minor": 0,
"metadata": {
"colab": {
"name": "WiC_self-stuffing_ExampleSizeComparison.ipynb",
"provenance": [],
"collapsed_sections": [],
"authorship_tag": "ABX9TyMfi/nyfCxWekDQ8kIyKUNO",
"include_colab_link": true
},
"kernelspec": {
"name": "python3",
"display_name": "Python 3"
}
},
"cells": [
{
"cell_type": "markdown",
"metadata": {
"id": "view-in-github",
"colab_type": "text"
},
"source": [
"<a href=\"https://colab.research.google.com/gist/brockmanmatt/0008cf6ae0fea4cb104111472012b864/wic_self-stuffing_examplesizecomparison.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
]
},
{
"cell_type": "code",
"metadata": {
"id": "J7wnsgT2kPut",
"colab_type": "code",
"colab": {
"resources": {
"http://localhost:8080/nbextensions/google.colab/files.js": {
"data": "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",
"ok": true,
"headers": [
[
"content-type",
"application/javascript"
]
],
"status": 200,
"status_text": ""
}
},
"base_uri": "https://localhost:8080/",
"height": 89
},
"outputId": "be181b0e-e7a7-4303-9c76-e6112498c898"
},
"source": [
"from google.colab import files\n",
"uploaded = files.upload()\n",
"print(\"done\")"
],
"execution_count": null,
"outputs": [
{
"output_type": "display_data",
"data": {
"text/html": [
"\n",
" <input type=\"file\" id=\"files-67e6c7a6-0fb4-4c59-a4e1-6ae43cb0efcf\" name=\"files[]\" multiple disabled\n",
" style=\"border:none\" />\n",
" <output id=\"result-67e6c7a6-0fb4-4c59-a4e1-6ae43cb0efcf\">\n",
" Upload widget is only available when the cell has been executed in the\n",
" current browser session. Please rerun this cell to enable.\n",
" </output>\n",
" <script src=\"/nbextensions/google.colab/files.js\"></script> "
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {
"tags": []
}
},
{
"output_type": "stream",
"text": [
"Saving key.json to key.json\n",
"done\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "WHPHrUnhpKnI",
"colab_type": "text"
},
"source": [
"I'll install the API"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "h4BX5shF5urh",
"colab_type": "text"
},
"source": [
"get wic dataset"
]
},
{
"cell_type": "code",
"metadata": {
"id": "zq0ltp2xn4yt",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 292
},
"outputId": "e6634189-d3e1-4748-c386-59ab1dd66b7b"
},
"source": [
"!pip install openai\n",
"import openai, json, pandas as pd, numpy as np, random"
],
"execution_count": null,
"outputs": [
{
"output_type": "stream",
"text": [
"Collecting openai\n",
"\u001b[?25l Downloading https://files.pythonhosted.org/packages/a8/65/c7461f4c87984534683f480ea5742777bc39bbf5721123194c2d0347dc1f/openai-0.2.4.tar.gz (157kB)\n",
"\r\u001b[K |██ | 10kB 10.3MB/s eta 0:00:01\r\u001b[K |████▏ | 20kB 2.2MB/s eta 0:00:01\r\u001b[K |██████▎ | 30kB 2.8MB/s eta 0:00:01\r\u001b[K |████████▍ | 40kB 3.1MB/s eta 0:00:01\r\u001b[K |██████████▍ | 51kB 2.4MB/s eta 0:00:01\r\u001b[K |████████████▌ | 61kB 2.7MB/s eta 0:00:01\r\u001b[K |██████████████▋ | 71kB 3.0MB/s eta 0:00:01\r\u001b[K |████████████████▊ | 81kB 3.3MB/s eta 0:00:01\r\u001b[K |██████████████████▊ | 92kB 3.4MB/s eta 0:00:01\r\u001b[K |████████████████████▉ | 102kB 3.4MB/s eta 0:00:01\r\u001b[K |███████████████████████ | 112kB 3.4MB/s eta 0:00:01\r\u001b[K |█████████████████████████ | 122kB 3.4MB/s eta 0:00:01\r\u001b[K |███████████████████████████ | 133kB 3.4MB/s eta 0:00:01\r\u001b[K |█████████████████████████████▏ | 143kB 3.4MB/s eta 0:00:01\r\u001b[K |███████████████████████████████▎| 153kB 3.4MB/s eta 0:00:01\r\u001b[K |████████████████████████████████| 163kB 3.4MB/s \n",
"\u001b[?25hRequirement already satisfied: requests>=2.20 in /usr/local/lib/python3.6/dist-packages (from openai) (2.23.0)\n",
"Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.6/dist-packages (from requests>=2.20->openai) (3.0.4)\n",
"Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.6/dist-packages (from requests>=2.20->openai) (1.24.3)\n",
"Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.6/dist-packages (from requests>=2.20->openai) (2020.6.20)\n",
"Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.6/dist-packages (from requests>=2.20->openai) (2.10)\n",
"Building wheels for collected packages: openai\n",
" Building wheel for openai (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
" Created wheel for openai: filename=openai-0.2.4-cp36-none-any.whl size=170709 sha256=cfc5e50e579a813bd45c5a9a087eea3547d042cc31f7ea8b93b18dbe7d3fe1ea\n",
" Stored in directory: /root/.cache/pip/wheels/74/96/c8/c6e170929c276b836613e1b9985343b501fe455e53d85e7d48\n",
"Successfully built openai\n",
"Installing collected packages: openai\n",
"Successfully installed openai-0.2.4\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "k67w5H0fpTkT",
"colab_type": "text"
},
"source": [
"Default keyword arguments to pass the aPI"
]
},
{
"cell_type": "code",
"metadata": {
"id": "e1EwpqqJkTYh",
"colab_type": "code",
"colab": {}
},
"source": [
"#arguments to send the API\n",
"kwargs = {\n",
"\"engine\":\"davinci\",\n",
"\"temperature\":0,\n",
"\"max_tokens\":200,\n",
"\"stop\":\"\\n\",\n",
"}\n",
"\n",
"kwargs2 = {\n",
"\"engine\":\"davinci\",\n",
"\"temperature\":0,\n",
"\"max_tokens\":200,\n",
"\"stop\":\"\\n\\n\",\n",
"}\n",
"\n",
"kwargs2Long = {\n",
"\"engine\":\"davinci\",\n",
"\"temperature\":0,\n",
"\"max_tokens\":500,\n",
"\"stop\":\"\\n\\n\",\n",
"}\n"
],
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "XPqahGflR_Tc",
"colab_type": "code",
"colab": {}
},
"source": [
"openai.api_key = json.load(open(\"key.json\", \"r\"))[\"key\"]"
],
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "sXTDJx0An9Bl",
"colab_type": "code",
"colab": {}
},
"source": [
"import datetime\n",
"def queryTwoLineLong(prompt, myKwargs = kwargs2Long):\n",
" \"\"\"\n",
" wrapper for the API to save the prompt and the result\n",
" \"\"\"\n",
" r = openai.Completion.create(prompt=prompt, **myKwargs)[\"choices\"][0][\"text\"].strip()\n",
" return r\n",
"\n",
"def queryTwoLine(prompt, myKwargs = kwargs2):\n",
" \"\"\"\n",
" wrapper for the API to save the prompt and the result\n",
" \"\"\"\n",
" r = openai.Completion.create(prompt=prompt, **myKwargs)[\"choices\"][0][\"text\"].strip()\n",
" return r\n",
"\n",
"def queryOneLine(prompt, myKwargs = kwargs):\n",
" \"\"\"\n",
" wrapper for the API to save the prompt and the result\n",
" \"\"\"\n",
" r = openai.Completion.create(prompt=prompt, **myKwargs)[\"choices\"][0][\"text\"].strip()\n",
" return r"
],
"execution_count": null,
"outputs": []
},
{
"cell_type": "markdown",
"metadata": {
"id": "EdFXafcJpZ3Q",
"colab_type": "text"
},
"source": [
"Test to make sure my query works"
]
},
{
"cell_type": "code",
"metadata": {
"id": "4SlyKgjyopPn",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 35
},
"outputId": "1cdf1920-81ad-4b9c-b8c5-3bf760b22615"
},
"source": [
"newKwargs = kwargs.copy()\n",
"newKwargs[\"stop\"] = \"\\n\"\n",
"queryOneLine(\"q: what is 1+1?\\na:\")"
],
"execution_count": null,
"outputs": [
{
"output_type": "execute_result",
"data": {
"application/vnd.google.colaboratory.intrinsic+json": {
"type": "string"
},
"text/plain": [
"'2'"
]
},
"metadata": {
"tags": []
},
"execution_count": 6
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "XDy96ovI9hJm",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 204
},
"outputId": "35666ab9-762c-4eb7-bd37-a355a474bc66"
},
"source": [
"!wget https://pilehvar.github.io/wic/package/WiC_dataset.zip"
],
"execution_count": null,
"outputs": [
{
"output_type": "stream",
"text": [
"--2020-08-04 14:26:54-- https://pilehvar.github.io/wic/package/WiC_dataset.zip\n",
"Resolving pilehvar.github.io (pilehvar.github.io)... 185.199.111.153, 185.199.110.153, 185.199.108.153, ...\n",
"Connecting to pilehvar.github.io (pilehvar.github.io)|185.199.111.153|:443... connected.\n",
"HTTP request sent, awaiting response... 200 OK\n",
"Length: 275513 (269K) [application/zip]\n",
"Saving to: ‘WiC_dataset.zip’\n",
"\n",
"\rWiC_dataset.zip 0%[ ] 0 --.-KB/s \rWiC_dataset.zip 100%[===================>] 269.06K --.-KB/s in 0.05s \n",
"\n",
"2020-08-04 14:26:54 (5.11 MB/s) - ‘WiC_dataset.zip’ saved [275513/275513]\n",
"\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "Gx-SlrBpSVXJ",
"colab_type": "code",
"colab": {}
},
"source": [
"import zipfile\n",
"with zipfile.ZipFile(\"WiC_dataset.zip\",\"r\") as zip_ref:\n",
" zip_ref.extractall(\".\")\n"
],
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"colab_type": "code",
"id": "CAHaOQKx9ogG",
"colab": {}
},
"source": [
"train = pd.read_csv(\"train/train.data.txt\", sep='\\t', header=None)\n",
"train.columns = [\"target\", \"pos\", \"position\", \"context-1\", \"context-2\"]\n",
"train_gold = pd.read_csv(\"train/train.gold.txt\", sep='\\t', header=None)\n",
"train_gold.columns = [\"label\"]\n",
"train = pd.concat([train_gold,train], axis=1)"
],
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "XC1pGirRwF8Z",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 669
},
"outputId": "0a038b69-5007-4f9a-876e-dff5633ba478"
},
"source": [
"train[train.pos==\"V\"].tail(20)"
],
"execution_count": null,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>label</th>\n",
" <th>target</th>\n",
" <th>pos</th>\n",
" <th>position</th>\n",
" <th>context-1</th>\n",
" <th>context-2</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>5378</th>\n",
" <td>T</td>\n",
" <td>meet</td>\n",
" <td>V</td>\n",
" <td>4-3</td>\n",
" <td>The company agrees to meet the cost of any rep...</td>\n",
" <td>Does this paper meet the requirements for the ...</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5384</th>\n",
" <td>T</td>\n",
" <td>anger</td>\n",
" <td>V</td>\n",
" <td>1-1</td>\n",
" <td>You anger too easily .</td>\n",
" <td>He angers easily .</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5386</th>\n",
" <td>F</td>\n",
" <td>admit</td>\n",
" <td>V</td>\n",
" <td>3-1</td>\n",
" <td>The French doors admit onto the yard .</td>\n",
" <td>He admitted his errors .</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5387</th>\n",
" <td>F</td>\n",
" <td>exhaust</td>\n",
" <td>V</td>\n",
" <td>0-4</td>\n",
" <td>Exhaust one 's savings .</td>\n",
" <td>This kind of work exhausts me .</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5389</th>\n",
" <td>F</td>\n",
" <td>kill</td>\n",
" <td>V</td>\n",
" <td>0-2</td>\n",
" <td>Kill the engine .</td>\n",
" <td>She was killed in the collision of three cars .</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5390</th>\n",
" <td>T</td>\n",
" <td>admit</td>\n",
" <td>V</td>\n",
" <td>1-1</td>\n",
" <td>To admit a serious thought into the mind .</td>\n",
" <td>She admitted us here .</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5394</th>\n",
" <td>F</td>\n",
" <td>write</td>\n",
" <td>V</td>\n",
" <td>6-1</td>\n",
" <td>How many books did Georges Simenon write ?</td>\n",
" <td>Please write to me every week .</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5395</th>\n",
" <td>T</td>\n",
" <td>enable</td>\n",
" <td>V</td>\n",
" <td>3-2</td>\n",
" <td>This skill will enable you to find a job on Wa...</td>\n",
" <td>The rope enables you to secure yourself when y...</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5397</th>\n",
" <td>T</td>\n",
" <td>answer</td>\n",
" <td>V</td>\n",
" <td>3-2</td>\n",
" <td>The man must answer to his employer for the mo...</td>\n",
" <td>She must answer for her actions .</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5400</th>\n",
" <td>F</td>\n",
" <td>hit</td>\n",
" <td>V</td>\n",
" <td>0-1</td>\n",
" <td>Hit the bottle .</td>\n",
" <td>He hit a home run .</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5403</th>\n",
" <td>T</td>\n",
" <td>sanitize</td>\n",
" <td>V</td>\n",
" <td>0-0</td>\n",
" <td>Sanitize the language in a book .</td>\n",
" <td>Sanitize history .</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5404</th>\n",
" <td>F</td>\n",
" <td>try</td>\n",
" <td>V</td>\n",
" <td>2-0</td>\n",
" <td>You are trying my patience .</td>\n",
" <td>Try the yak butter .</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5405</th>\n",
" <td>T</td>\n",
" <td>classify</td>\n",
" <td>V</td>\n",
" <td>3-1</td>\n",
" <td>How would you classify these pottery shards --...</td>\n",
" <td>She classified the works as ' dangerous ' .</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5407</th>\n",
" <td>F</td>\n",
" <td>break</td>\n",
" <td>V</td>\n",
" <td>8-1</td>\n",
" <td>My daughter 's fancy wedding is going to break...</td>\n",
" <td>He broke the glass plate .</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5413</th>\n",
" <td>F</td>\n",
" <td>render</td>\n",
" <td>V</td>\n",
" <td>0-0</td>\n",
" <td>Render thanks .</td>\n",
" <td>Render fat in a casserole .</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5415</th>\n",
" <td>T</td>\n",
" <td>drive</td>\n",
" <td>V</td>\n",
" <td>1-1</td>\n",
" <td>I drive to work every day .</td>\n",
" <td>We drove to the university every morning .</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5417</th>\n",
" <td>T</td>\n",
" <td>keep</td>\n",
" <td>V</td>\n",
" <td>0-0</td>\n",
" <td>Keep open the possibility of a merger .</td>\n",
" <td>Keep my seat , please .</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5419</th>\n",
" <td>T</td>\n",
" <td>stretch</td>\n",
" <td>V</td>\n",
" <td>1-0</td>\n",
" <td>Cats stretch with equal ease and agility beyon...</td>\n",
" <td>Stretch your legs !</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5425</th>\n",
" <td>T</td>\n",
" <td>answer</td>\n",
" <td>V</td>\n",
" <td>0-0</td>\n",
" <td>Answer the riddle .</td>\n",
" <td>Answer a question .</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5426</th>\n",
" <td>T</td>\n",
" <td>play</td>\n",
" <td>V</td>\n",
" <td>0-0</td>\n",
" <td>Play the casinos in Trouville .</td>\n",
" <td>Play the races .</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" label ... context-2\n",
"5378 T ... Does this paper meet the requirements for the ...\n",
"5384 T ... He angers easily .\n",
"5386 F ... He admitted his errors .\n",
"5387 F ... This kind of work exhausts me .\n",
"5389 F ... She was killed in the collision of three cars .\n",
"5390 T ... She admitted us here .\n",
"5394 F ... Please write to me every week .\n",
"5395 T ... The rope enables you to secure yourself when y...\n",
"5397 T ... She must answer for her actions .\n",
"5400 F ... He hit a home run .\n",
"5403 T ... Sanitize history .\n",
"5404 F ... Try the yak butter .\n",
"5405 T ... She classified the works as ' dangerous ' .\n",
"5407 F ... He broke the glass plate .\n",
"5413 F ... Render fat in a casserole .\n",
"5415 T ... We drove to the university every morning .\n",
"5417 T ... Keep my seat , please .\n",
"5419 T ... Stretch your legs !\n",
"5425 T ... Answer a question .\n",
"5426 T ... Play the races .\n",
"\n",
"[20 rows x 6 columns]"
]
},
"metadata": {
"tags": []
},
"execution_count": 35
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "28-_k4pT6psq",
"colab_type": "code",
"colab": {}
},
"source": [
"for row in train[train.pos==\"V\"].tail(20).iterrows():\n",
" print(\"Contexts: '{}'; '{}'\".format(row[1][\"context-1\"], row[1][\"context-2\"]))\n",
" print(\"Term: '{}'\".format(row[1][\"target\"]))\n",
" print(\"Meaning:\")\n",
" if row[1][\"label\"] == \"T\":\n",
" print(\"They are similar\")\n",
" else:\n",
" print(\"They are dissimilar\")\n"
],
"execution_count": null,
"outputs": []
},
{
"cell_type": "markdown",
"metadata": {
"id": "TDlFJMvB636g",
"colab_type": "text"
},
"source": [
""
]
},
{
"cell_type": "code",
"metadata": {
"id": "Eo29eYgMv-M0",
"colab_type": "code",
"colab": {}
},
"source": [
"fewShotVerb = \"\"\"Contexts: 'The French doors admit onto the yard .'; 'He admitted his errors .'\n",
"Term: 'admit'\n",
"Meaning: In the first sentence, 'admit' means to provide passage. In the second, 'admit' means to take responsibility.\n",
"They are dissimilar\n",
"\n",
"Contexts: 'The company agrees to meet the cost of any repairs .'; 'Does this paper meet the requirements for the degree ?'\n",
"Term: 'meet'\n",
"Meaning: In the first sentence, 'meet' means to fulfill. In the second, 'meet' means to take fulfill.\n",
"They are similar\n",
"\n",
"Contexts: 'You anger too easily .'; 'He angers easily .'\n",
"Term: 'anger'\n",
"Meaning: In the first sentence, 'anger' means to enrage. In the second, 'anger' means to enrage.\n",
"They are similar\n",
"\n",
"Contexts: 'Exhaust one 's savings .'; 'This kind of work exhausts me .'\n",
"Term: 'exhaust'\n",
"Meaning: In the first sentence, 'exhaust' means to empty. In the second, 'exhaust' means to tire.\n",
"They are dissimilar\n",
"\n",
"Contexts: 'Kill the engine .'; 'She was killed in the collision of three cars .'\n",
"Term: 'kill'\n",
"Meaning: In the first sentence, 'kill' means to turn off. In the second, 'kill' refers to dying.\n",
"They are dissimilar\n",
"\n",
"Contexts: 'To admit a serious thought into the mind .'; 'She admitted us here .'\n",
"Term: 'admit'\n",
"Meaning: In the first sentence, 'admit' means to allow in. In the second, 'admit' means to allow in.\n",
"They are similar\n",
"\n",
"Contexts: 'How many books did Georges Simenon write ?'; 'Please write to me every week .'\n",
"Term: 'write'\n",
"Meaning: In the first sentence, 'write' means to publish. In the second, 'write' means to communicate with. \n",
"They are dissimilar\n",
"\n",
"Contexts: 'This skill will enable you to find a job on Wall Street .'; 'The rope enables you to secure yourself when you climb the mountain .'\n",
"Term: 'enable'\n",
"Meaning: In the first sentence, 'enable' means to give an ability. In the second, 'enable' means to give an ability.\n",
"They are similar\n",
"\n",
"Contexts: 'The man must answer to his employer for the money entrusted to his care .'; 'She must answer for her actions .'\n",
"Term: 'answer'\n",
"Meaning: In the first sentence, 'answer' means to take responsibility. In the second, 'answer' means to take responsibility.\n",
"They are similar\n",
"\n",
"Contexts: 'Hit the bottle .'; 'He hit a home run .'\n",
"Term: 'hit'\n",
"Meaning: In the first sentence, 'hit' means to use. In the second, 'hit' means to strike.\n",
"They are dissimilar\n",
"\n",
"Contexts: 'Sanitize the language in a book .'; 'Sanitize history .'\n",
"Term: 'Sanitize'\n",
"Meaning: In the first sentence, 'sanitize' means edit. In the second 'sanitize' means to edit.\n",
"They are similar\n",
"\n",
"Contexts: 'You are trying my patience .'; 'Try the yak butter .'\n",
"Term: 'try'\n",
"Meaning: In the first sentence, 'try' means to test something. In the second 'try' means to taste.\n",
"They are dissimilar\n",
"\n",
"Contexts: 'I drive to work every day .'; 'We drove to the university every morning .'\n",
"Term: 'drive'\n",
"Meaning: In the first sentence, 'drive' means to take a vehicle. In the second 'drive' means to take a vehicle.\n",
"They are similar\n",
"\n",
"Contexts: 'My daughter 's fancy wedding is going to break me .'; 'He broke the glass plate .'\n",
"Terrm: 'break'\n",
"Meaning: In the first sentence, 'break' means to bankrupt. In the second 'break' means to shatter someting.\n",
"They are dissimilar\n",
"\n",
"Contexts: 'How many books did Georges Simenon write .'; 'Please write to me every week .'\n",
"Term: 'write'\n",
"Meaning: In the first sentence, 'write' refers to publishing. In the second 'write' means to correspond.\n",
"They are dissimilar\n",
"\n",
"Contexts: 'Keep my seat , please'; 'Keep open the possibility of a merger .'\n",
"Term: 'keep'\n",
"Meaning: In the first sentence, 'keep' refers to leaving open. In the second 'keep' means leaving open.\n",
"They are similar\n",
"\n",
"Contexts: 'Cats stretch with equal ease and agility beyond the point that breaks a man on the rack .'; 'Stretch your legs !'\n",
"Term: 'stretch'\n",
"Meaning: In the first sentence, 'stretch' refers to extending body parts. In the second 'stretch' means extending body parts.\n",
"They are similar\n",
"\n",
"Contexts: 'Answer the riddle .'; 'Answer a question .'\n",
"Term: 'answer'\n",
"Meaning: In the first sentence, 'answer' means to respond. In the second 'answer' means to respond.\n",
"They are similar\n",
"\n",
"Contexts: 'Play the casinos in Trouville .'; 'Play the races .'\n",
"Term: 'play'\n",
"Meaning: In the first sentence, 'play' means to gamble. In the second 'play' means to gamble.\n",
"They are similar\"\"\".split(\"\\n\\n\")"
],
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "5mYRyWJD3_9v",
"colab_type": "code",
"colab": {}
},
"source": [
"def queryVerb(row, numExamples):\n",
" prefix = \"Does each pair of sentences uses the selected term in a similar sense?\\n\\n\"\n",
"\n",
" context = \"\"\"Does each pair of sentences uses the selected term in a similar sense?\n",
"\n",
"Contexts: '{}'; '{}'\n",
"Term: '{}'\n",
"Meaning:\"\"\".format(row[1][\"context-1\"], row[1][\"context-2\"], row[1][\"target\"])\n",
" return queryTwoLine(prefix + \"\\n\\n\".join(fewShotVerb[:numExamples]) + \"\\n\\n\" + context)"
],
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "FZULhAAIGzTL",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 85
},
"outputId": "a9a8c3eb-30dd-426f-9911-f6b90fed49f0"
},
"source": [
"print(\"\\n\\n\".join(fewShotVerb[:1]))"
],
"execution_count": null,
"outputs": [
{
"output_type": "stream",
"text": [
"Contexts: 'The French doors admit onto the yard .'; 'He admitted his errors .'\n",
"Term: 'admit'\n",
"Meaning: In the first sentence, 'admit' means to provide passage. In the second, 'admit' means to take responsibility.\n",
"They are dissimilar\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "_w8z-m1LEppI",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "46f95245-105b-4b50-a63f-80391cc301f2"
},
"source": [
"len(fewShotVerb)"
],
"execution_count": null,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"19"
]
},
"metadata": {
"tags": []
},
"execution_count": 67
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "aavDvEwiHV6D",
"colab_type": "code",
"colab": {}
},
"source": [
"scores = pd.DataFrame()"
],
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "VZsZ1DqqwoSb",
"colab_type": "code",
"colab": {}
},
"source": [
"for i in range(7,12):\n",
"\n",
" print(i)\n",
" scores.at[i, \"correct\"] = 0\n",
" scores.at[i, \"completed\"] = 0\n",
"\n",
" for row in train[train.pos==\"V\"].head(5).iterrows():\n",
" print(row[1][\"context-1\"], row[1][\"context-2\"])\n",
" actual = row[1][\"label\"]\n",
" output = (queryVerb(row,i))\n",
" scores.at[i, \"completed\"] +=1\n",
"\n",
" if actual == \"T\":\n",
" if output.strip().split()[-1]==(\"similar\"):\n",
" correct += 1\n",
" scores.at[i, \"correct\"] += 1\n",
" if actual == \"F\":\n",
" if output.strip().split()[-1]==(\"dissimilar\"):\n",
" scores.at[i, \"correct\"] +=1 \n",
"\n",
"\n",
" print(output)"
],
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "vh_TbqLeGpm8",
"colab_type": "code",
"colab": {}
},
"source": [
"scores"
],
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "VqD89fdR0Mou",
"colab_type": "code",
"colab": {}
},
"source": [
"pd.set_option('display.max_colwidth', -1)\n"
],
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "nNkkgtZVcA7x",
"colab_type": "code",
"colab": {}
},
"source": [
"for row in train[train.pos==\"N\"].tail(20).iterrows():\n",
" print(\"Contexts: '{}'; '{}'\".format(row[1][\"context-1\"], row[1][\"context-2\"]))\n",
" print(\"Term: '{}'\".format(row[1][\"target\"]))\n",
" print(\"Meaning: In the first sentence, '{}' means ___; in the second '{}' means ____\".format(row[1][\"target\"],row[1][\"target\"]))\n",
" if row[1][\"label\"] == \"T\":\n",
" print(\"They are similar\")\n",
" else:\n",
" print(\"They are dissimilar\")\n",
" print()\n"
],
"execution_count": null,
"outputs": []
},
{
"cell_type": "markdown",
"metadata": {
"id": "FY-EdpDety9V",
"colab_type": "text"
},
"source": [
"# I swapped 1 and 2 so it starts with dissimilar; not sure why this is better."
]
},
{
"cell_type": "code",
"metadata": {
"id": "OqEa3PZUyd8a",
"colab_type": "code",
"colab": {}
},
"source": [
"fewShotNoun = \"\"\"Contexts: 'A history of France .'; 'A critical time in the school 's history .'\n",
"Term: 'history'\n",
"Meaning: In the first sentence, 'history' refers to a record; in the second it means past.\n",
"The meanings are dissimilar\n",
"\n",
"Contexts: 'I do it for the fun of it .'; 'He is fun to have around .'\n",
"Term: 'fun'\n",
"Meaning: In the first sentence, 'fun' means having pleasure; in the second it also means pleasurable.\n",
"The meanings are similar\n",
"\n",
"Contexts: 'The rate of production at the factory is skyrocketing .'; 'He works at a great rate .'\n",
"Term: 'rate'\n",
"Meaning: In the first sentence, 'rate' refers to speed; in the second it also means speed.\n",
"The meanings are similar\n",
"\n",
"Contexts: 'Get to the point .'; 'At that point I had to leave .'\n",
"Term: 'point'\n",
"Meaning: In the first sentence, 'point' refers to a topic; in the second it means time.\n",
"The meanings are dissimilar\n",
"\n",
"Contexts: 'Kronas kurss — the exchange rate of the krona .'; 'Grāmata maksā piecas kronas — the book costs five krona .'\n",
"Term: 'krona'\n",
"Meaning: In the first sentence, 'krona' means money; in the second it also means money.\n",
"The meanings are similar\n",
"\n",
"Contexts: 'Armored from head to foot .'; 'The swiftest of foot .'\n",
"Term: 'foot'\n",
"Meaning: In the first sentence, 'foot' refers to a body part; in the second it means speed.\n",
"The meanings are dissimilar\n",
"\n",
"Contexts: 'A patch of clouds .'; 'Patches of thin ice .'\n",
"Term: 'patch'\n",
"Meaning: In the first sentence, 'patch' means cluster; in the second it also means cluster.\n",
"The meanings are similar\n",
"\n",
"Contexts: 'The misery and wretchedness of those slums is intolerable .'; 'She was exhausted by her misery and grief .'\n",
"Term: 'misery'\n",
"Meaning: In the first sentence, 'misery' means a state of unhappiness; in the second it refers to a feeling of unhappiness.\n",
"The meanings are dissimilar\n",
"\n",
"Contexts: 'Women carrying home shopping did n't give me a second glance .'; 'On Saturdays we usually do the shopping .'\n",
"Term: 'shopping'\n",
"Meaning: In the first sentence, 'shopping' refers to purchases; in the second it means buying things.\n",
"The meanings are dissimilar\n",
"\n",
"Contexts: 'While being impulsive can be great for artists , it is not a desirable quality for engineers .'; 'Security , stability , and efficiency are good qualities of an operating system .'\n",
"Term: 'quality'\n",
"Meaning: In the first sentence, 'quality' means attribute; in the second it also means attribute.\n",
"The meanings are similar\n",
"\n",
"Contexts: 'The cinema relies on apparent motion .'; 'He made a motion to adjourn .'\n",
"Term: 'motion'\n",
"Meaning: In the first sentence, 'motion' means movement; in the second it means proposal.\n",
"The meanings are dissimilar\n",
"\n",
"Contexts: 'It vanished into the night .'; 'The cat disappeared into the night .'\n",
"Term: 'night'\n",
"Meaning: In the first sentence, 'night' means darkness; in the second it also means darkness.\n",
"The meanings are similar\n",
"\n",
"Contexts: 'He threw the ball into the air .'; 'A smell of chemicals in the air .'\n",
"Term: 'air'\n",
"Meaning: In the first sentence, 'air' means sky; in the second it means a gas.\n",
"The meanings are dissimilar\n",
"\n",
"Contexts: 'Those clouds show little sign of raining soon .'; 'Signs of disease are objective , whereas symptoms are subjective .'\n",
"Term: 'sign'\n",
"Meaning: In the first sentence, 'sign' means indication; in the second it also means indication.\n",
"The meanings are similar\n",
"\n",
"Contexts: 'We added a new rosebush to our rose bed .'; 'The gardener planted a bed of roses .'\n",
"Term: 'bed'\n",
"Meaning: In the first sentence, 'bed' means garden; in the second it also means garden.\n",
"The meanings are similar\n",
"\n",
"Contexts: 'His state of health .'; 'In a weak financial state .'\n",
"Term: 'state'\n",
"Meaning: In the first sentence, 'state' means condition; in the second it also means condition.\n",
"The meanings are similar\n",
"\n",
"Contexts: 'Likes a drink before dinner .'; 'Can I buy you a drink ?'\n",
"Term: 'drink'\n",
"Meaning: In the first sentence, 'drink' means beverage; in the second it also means beverage.\n",
"The meanings are similar\n",
"\n",
"Contexts: 'Piecas kronas — five krona .'; 'Kronas kurss — the exchange rate of the krona .'\n",
"Term: 'krona'\n",
"Meaning: In the first sentence, 'krona' means money; in the second it also means money.\n",
"The meanings are similar\n",
"\n",
"Contexts: 'The harder the conflict the more glorious the triumph \"-- Thomas Paine .'; 'The conflict between the government and the rebels began three years ago .'\n",
"Term: 'conflict'\n",
"Meaning: In the first sentence, 'conflict' means struggle; in the second it also means struggle.\n",
"The meanings are similar\n",
"\n",
"Contexts: 'An invasion of bees .'; 'An invasion of mobile phones .'\n",
"Term: 'invasion'\n",
"Meaning: In the first sentence, 'invasion' means numeric growth; in the second it also means numeric growth.\n",
"The meanings are similar\"\"\".split(\"\\n\\n\")"
],
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "0h7fz2EV3xhF",
"colab_type": "code",
"colab": {}
},
"source": [
"def queryNoun(row, numExamples):\n",
" prefix = \"Does each pair of sentences uses the selected term in a similar sense?\\n\\n\"\n",
"\n",
" context = \"\"\"Does each pair of sentences uses the selected term in a similar sense?\n",
" \n",
"Contexts: '{}'; '{}'\n",
"Term: '{}'\n",
"Meaning:\"\"\".format(row[1][\"context-1\"], row[1][\"context-2\"], row[1][\"target\"])\n",
" return queryTwoLine(prefix + \"\\n\\n\".join(fewShotNoun[:numExamples]) + \"\\n\\n\" + context)"
],
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "3mt98W_QwwC9",
"colab_type": "code",
"colab": {}
},
"source": [
"scores = pd.DataFrame()\n",
"for i in range(4,12):\n",
"\n",
" print(i)\n",
" scores.at[i, \"correct\"] = 0\n",
" scores.at[i, \"completed\"] = 0\n",
"\n",
" for row in train[train.pos==\"N\"].head(2).iterrows():\n",
" print(row[1][\"context-1\"], row[1][\"context-2\"])\n",
" actual = row[1][\"label\"]\n",
" output = (queryNoun(row,i))\n",
" scores.at[i, \"completed\"] +=1\n",
"\n",
" if actual == \"T\":\n",
" if output.strip().split()[-1]==(\"similar\"):\n",
" correct += 1\n",
" scores.at[i, \"correct\"] += 1\n",
" if actual == \"F\":\n",
" if output.strip().split()[-1]==(\"dissimilar\"):\n",
" scores.at[i, \"correct\"] +=1 \n",
"\n",
"\n",
" print(output)"
],
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "65U-5Nwbi8Y5",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 111
},
"outputId": "da98696d-20c7-447e-c8ea-c8c0acc1ca55"
},
"source": [
"scores"
],
"execution_count": null,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>correct</th>\n",
" <th>completed</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>5</th>\n",
" <td>4.0</td>\n",
" <td>5.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>6</th>\n",
" <td>1.0</td>\n",
" <td>2.0</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" correct completed\n",
"5 4.0 5.0\n",
"6 1.0 2.0"
]
},
"metadata": {
"tags": []
},
"execution_count": 132
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "WU5RiResySom",
"colab_type": "code",
"colab": {}
},
"source": [
"dev = pd.read_csv(\"dev/dev.data.txt\", sep='\\t', header=None)\n",
"dev.columns = [\"target\", \"pos\", \"position\", \"context-1\", \"context-2\"]\n",
"dev_gold = pd.read_csv(\"dev/dev.gold.txt\", sep='\\t', header=None)\n",
"dev_gold.columns = [\"label\"]\n",
"dev = pd.concat([dev_gold,dev], axis=1)"
],
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "KtWXt5M54bZm",
"colab_type": "code",
"colab": {}
},
"source": [
"devResults = {}"
],
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "UAufc6pP3qYj",
"colab_type": "code",
"colab": {}
},
"source": [
"for i in range(5,20):\n",
" if i not in devResults:\n",
" devResults[i] = {}\n",
" correct = 0\n",
" complete = 0\n",
" print(\"CHECKING {}\".format(i))\n",
" for row in dev.iterrows():\n",
"\n",
" if row[0] in devResults[i]:\n",
" continue\n",
"\n",
" q1 = row[1][\"context-1\"]\n",
" q2 = row[1][\"context-2\"]\n",
" target = row[1][\"target\"]\n",
" actual = row[1][\"label\"]\n",
"\n",
" if row[1][\"pos\"] == \"N\":\n",
" output = queryNoun(row, i)\n",
"\n",
" else:\n",
" output = queryVerb(row, i)\n",
"\n",
" myResults = {}\n",
" myResults[\"q1\"] = q1\n",
" myResults[\"q2\"] = q2\n",
"\n",
" myResults[\"pos\"] = row[1][\"pos\"]\n",
"\n",
" myResults[\"target\"] = target\n",
"\n",
" myResults[\"output\"] = output\n",
"\n",
" myResults[\"actual\"] = actual\n",
" devResults[i][row[0]] = myResults\n",
" complete +=1\n",
" if actual == \"T\":\n",
" if output.strip().split()[-1]==(\"similar\"):\n",
" correct += 1\n",
" if actual == \"F\":\n",
" if output.strip().split()[-1]==(\"dissimilar\"):\n",
" correct += 1\n",
" if row[0] %10 ==0:\n",
" print (\"Complete: {} Correct: {} Wrong: {}\".format(complete, correct, complete-correct))"
],
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "uV4UZ2MImGeH",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 17
},
"outputId": "b3ff3b88-bf55-4e1d-b07f-678e3501e620"
},
"source": [
"import pickle\n",
"with open(\"devResults_selfStuff_5-20_600.pkl\", \"wb\") as fh:\n",
" pickle.dump(devResults, fh)\n",
"files.download(\"devResults_selfStuff_5-20_600.pkl\")\n"
],
"execution_count": null,
"outputs": [
{
"output_type": "display_data",
"data": {
"application/javascript": [
"\n",
" async function download(id, filename, size) {\n",
" if (!google.colab.kernel.accessAllowed) {\n",
" return;\n",
" }\n",
" const div = document.createElement('div');\n",
" const label = document.createElement('label');\n",
" label.textContent = `Downloading \"${filename}\": `;\n",
" div.appendChild(label);\n",
" const progress = document.createElement('progress');\n",
" progress.max = size;\n",
" div.appendChild(progress);\n",
" document.body.appendChild(div);\n",
"\n",
" const buffers = [];\n",
" let downloaded = 0;\n",
"\n",
" const channel = await google.colab.kernel.comms.open(id);\n",
" // Send a message to notify the kernel that we're ready.\n",
" channel.send({})\n",
"\n",
" for await (const message of channel.messages) {\n",
" // Send a message to notify the kernel that we're ready.\n",
" channel.send({})\n",
" if (message.buffers) {\n",
" for (const buffer of message.buffers) {\n",
" buffers.push(buffer);\n",
" downloaded += buffer.byteLength;\n",
" progress.value = downloaded;\n",
" }\n",
" }\n",
" }\n",
" const blob = new Blob(buffers, {type: 'application/binary'});\n",
" const a = document.createElement('a');\n",
" a.href = window.URL.createObjectURL(blob);\n",
" a.download = filename;\n",
" div.appendChild(a);\n",
" a.click();\n",
" div.remove();\n",
" }\n",
" "
],
"text/plain": [
"<IPython.core.display.Javascript object>"
]
},
"metadata": {
"tags": []
}
},
{
"output_type": "display_data",
"data": {
"application/javascript": [
"download(\"download_2f20ae32-7616-47ae-9cd4-bfad021f4099\", \"devResults_selfStuff_5-20_600.pkl\", 1607424)"
],
"text/plain": [
"<IPython.core.display.Javascript object>"
]
},
"metadata": {
"tags": []
}
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "4ouRDBwLpQbd",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 680
},
"outputId": "ca88add2-5955-48a2-afa7-102302c5ff0b"
},
"source": [
"evaluations = pd.DataFrame()\n",
"falseTruePositives = pd.DataFrame()\n",
"for exampleSize in devResults:\n",
" devDf = pd.DataFrame(devResults[exampleSize]).T\n",
" if len(devDf) < 600:\n",
" continue\n",
" devDf[\"pred\"] = devDf[\"output\"].apply(lambda x: \"F\" if x.endswith(\"dissimilar\") else \"T\")\n",
" print(\"EXAMPLE SIZE: {}\".format(exampleSize))\n",
"\n",
" pct_1 = (devDf.pred ==devDf[\"actual\"]).mean()\n",
" print(\"accuracy comparing initial answers: {}\".format(np.round(100*pct_1)))\n",
" evaluations.at[exampleSize, \"Combined\"] = pct_1\n",
"\n",
" nouns = (devDf[devDf.pos==\"N\"].pred ==devDf[devDf.pos==\"N\"][\"actual\"]).mean()\n",
" evaluations.at[exampleSize, \"Nouns\"] = nouns\n",
"\n",
" verbs = (devDf[devDf.pos==\"V\"].pred ==devDf[devDf.pos==\"V\"][\"actual\"]).mean()\n",
" evaluations.at[exampleSize, \"verbs\"] = verbs\n",
" \n",
" print (len(devDf))\n",
"\n",
" verbPositive = (devDf[(devDf.pos==\"V\")&(devDf.actual==\"T\")].pred ==devDf[(devDf.pos==\"V\")&(devDf.actual==\"T\")][\"actual\"]).mean()\n",
" falseTruePositives.at[exampleSize, \"verbPositive\"] = verbPositive\n",
" verbNegative = (devDf[(devDf.pos==\"V\")&(devDf.actual==\"F\")].pred ==devDf[(devDf.pos==\"V\")&(devDf.actual==\"F\")][\"actual\"]).mean()\n",
" falseTruePositives.at[exampleSize, \"verbNegative\"] = verbNegative\n",
" \n",
" nounPositive = (devDf[(devDf.pos==\"N\")&(devDf.actual==\"T\")].pred ==devDf[(devDf.pos==\"N\")&(devDf.actual==\"T\")][\"actual\"]).mean()\n",
" falseTruePositives.at[exampleSize, \"nounPositive\"] = nounPositive\n",
" nounNegative = (devDf[(devDf.pos==\"N\")&(devDf.actual==\"F\")].pred ==devDf[(devDf.pos==\"N\")&(devDf.actual==\"F\")][\"actual\"]).mean()\n",
" falseTruePositives.at[exampleSize, \"nounNegative\"] = nounNegative\n",
"\n"
],
"execution_count": null,
"outputs": [
{
"output_type": "stream",
"text": [
"EXAMPLE SIZE: 5\n",
"accuracy comparing initial answers: 59.0\n",
"638\n",
"EXAMPLE SIZE: 6\n",
"accuracy comparing initial answers: 61.0\n",
"638\n",
"EXAMPLE SIZE: 7\n",
"accuracy comparing initial answers: 60.0\n",
"638\n",
"EXAMPLE SIZE: 8\n",
"accuracy comparing initial answers: 61.0\n",
"638\n",
"EXAMPLE SIZE: 9\n",
"accuracy comparing initial answers: 61.0\n",
"638\n",
"EXAMPLE SIZE: 10\n",
"accuracy comparing initial answers: 60.0\n",
"638\n",
"EXAMPLE SIZE: 11\n",
"accuracy comparing initial answers: 61.0\n",
"638\n",
"EXAMPLE SIZE: 12\n",
"accuracy comparing initial answers: 62.0\n",
"638\n",
"EXAMPLE SIZE: 13\n",
"accuracy comparing initial answers: 61.0\n",
"638\n",
"EXAMPLE SIZE: 14\n",
"accuracy comparing initial answers: 60.0\n",
"638\n",
"EXAMPLE SIZE: 15\n",
"accuracy comparing initial answers: 60.0\n",
"638\n",
"EXAMPLE SIZE: 16\n",
"accuracy comparing initial answers: 59.0\n",
"638\n",
"EXAMPLE SIZE: 17\n",
"accuracy comparing initial answers: 61.0\n",
"638\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "84EcbAvapQYb",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 452
},
"outputId": "33393b31-507e-4638-d0e0-d13a111f6b7a"
},
"source": [
"100*evaluations"
],
"execution_count": null,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>Combined</th>\n",
" <th>Nouns</th>\n",
" <th>verbs</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>5</th>\n",
" <td>58.777429</td>\n",
" <td>56.708861</td>\n",
" <td>62.139918</td>\n",
" </tr>\n",
" <tr>\n",
" <th>6</th>\n",
" <td>61.285266</td>\n",
" <td>60.759494</td>\n",
" <td>62.139918</td>\n",
" </tr>\n",
" <tr>\n",
" <th>7</th>\n",
" <td>59.874608</td>\n",
" <td>58.227848</td>\n",
" <td>62.551440</td>\n",
" </tr>\n",
" <tr>\n",
" <th>8</th>\n",
" <td>60.658307</td>\n",
" <td>60.253165</td>\n",
" <td>61.316872</td>\n",
" </tr>\n",
" <tr>\n",
" <th>9</th>\n",
" <td>60.658307</td>\n",
" <td>62.278481</td>\n",
" <td>58.024691</td>\n",
" </tr>\n",
" <tr>\n",
" <th>10</th>\n",
" <td>59.561129</td>\n",
" <td>57.974684</td>\n",
" <td>62.139918</td>\n",
" </tr>\n",
" <tr>\n",
" <th>11</th>\n",
" <td>60.971787</td>\n",
" <td>59.746835</td>\n",
" <td>62.962963</td>\n",
" </tr>\n",
" <tr>\n",
" <th>12</th>\n",
" <td>61.912226</td>\n",
" <td>62.025316</td>\n",
" <td>61.728395</td>\n",
" </tr>\n",
" <tr>\n",
" <th>13</th>\n",
" <td>60.501567</td>\n",
" <td>59.493671</td>\n",
" <td>62.139918</td>\n",
" </tr>\n",
" <tr>\n",
" <th>14</th>\n",
" <td>60.188088</td>\n",
" <td>58.481013</td>\n",
" <td>62.962963</td>\n",
" </tr>\n",
" <tr>\n",
" <th>15</th>\n",
" <td>59.717868</td>\n",
" <td>58.734177</td>\n",
" <td>61.316872</td>\n",
" </tr>\n",
" <tr>\n",
" <th>16</th>\n",
" <td>59.404389</td>\n",
" <td>57.468354</td>\n",
" <td>62.551440</td>\n",
" </tr>\n",
" <tr>\n",
" <th>17</th>\n",
" <td>60.815047</td>\n",
" <td>60.253165</td>\n",
" <td>61.728395</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" Combined Nouns verbs\n",
"5 58.777429 56.708861 62.139918\n",
"6 61.285266 60.759494 62.139918\n",
"7 59.874608 58.227848 62.551440\n",
"8 60.658307 60.253165 61.316872\n",
"9 60.658307 62.278481 58.024691\n",
"10 59.561129 57.974684 62.139918\n",
"11 60.971787 59.746835 62.962963\n",
"12 61.912226 62.025316 61.728395\n",
"13 60.501567 59.493671 62.139918\n",
"14 60.188088 58.481013 62.962963\n",
"15 59.717868 58.734177 61.316872\n",
"16 59.404389 57.468354 62.551440\n",
"17 60.815047 60.253165 61.728395"
]
},
"metadata": {
"tags": []
},
"execution_count": 181
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "6zcfKvtQriuJ",
"colab_type": "code",
"colab": {}
},
"source": [
"import matplotlib.pyplot as plt\n"
],
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "EfHxK33zpUOJ",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 421
},
"outputId": "a66198af-6196-4182-cc3b-726ed6342669"
},
"source": [
"fig, ax = plt.subplots(facecolor=\"w\", figsize=(8,6))\n",
"(100*evaluations).plot(title=\"Performance on Full WiC Dev by Number of Examples in Prompt\", ax=ax)\n",
"ax.set_ylabel(\"% accuracy\")\n",
"ax.set_xlabel(\"Number Examples\")\n"
],
"execution_count": null,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"Text(0.5, 0, 'Number Examples')"
]
},
"metadata": {
"tags": []
},
"execution_count": 188
},
{
"output_type": "display_data",
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 576x432 with 1 Axes>"
]
},
"metadata": {
"tags": []
}
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "gk-ZWFovpT4p",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 421
},
"outputId": "b7c46483-2f4d-4753-c7d1-a1cb13808c6c"
},
"source": [
"fig, ax = plt.subplots(facecolor=\"w\", figsize=(8,6))\n",
"#style = [\"|\"]\n",
"(100*falseTruePositives).plot(title=\"Performance on Full WiC Dev by Number of Examples in Prompt\", ax=ax, style=[\"-\", \":\", \"-\", \":\"], color=[\"blue\", \"blue\", \"green\", \"green\"])\n",
"ax.set_ylabel(\"% accuracy\")\n",
"ax.set_xlabel(\"Number Examples\")\n"
],
"execution_count": null,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"Text(0.5, 0, 'Number Examples')"
]
},
"metadata": {
"tags": []
},
"execution_count": 190
},
{
"output_type": "display_data",
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 576x432 with 1 Axes>"
]
},
"metadata": {
"tags": []
}
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "2m-oLeAkTf-x",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "b1dc09ca-7b8f-4af2-8734-d9728a9945c6"
},
"source": [
""
],
"execution_count": null,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"638"
]
},
"metadata": {
"tags": []
},
"execution_count": 205
}
]
}
]
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment