File size: 5,593 Bytes
96b6673
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "5e21c7be",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package punkt to\n",
      "[nltk_data]     /Users/shenjiajun/nltk_data...\n",
      "[nltk_data]   Package punkt is already up-to-date!\n",
      "/Users/shenjiajun/miniconda3/envs/torch_gpu/lib/python3.10/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n"
     ]
    }
   ],
   "source": [
    "from citekit.cite_modules.LLM import LLM\n",
    "from citekit.cite_modules.augment_model import AttributingModule\n",
    "from citekit.pipeline.pipeline import Sequence\n",
    "from citekit.prompt.prompt import Prompt\n",
    "from citekit.Dataset.Dataset import FileDataset\n",
    "from citekit.evaluator.evaluator import DefaultEvaluator\n",
    "import json"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0260d8c5",
   "metadata": {},
   "source": [
    "# Data Loading"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "c374b212",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Instruction: Write an accurate, engaging, and concise answer for the given question using only the provided search results (some of which might be irrelevant) and cite them properly. Use an unbiased and journalistic tone. Always cite for any factual claim. When citing several search results, use [1][2][3]. Cite at least one document and at most three documents in each sentence. If multiple documents support the sentence, only cite a minimum sufficient subset of the documents.\n"
     ]
    }
   ],
   "source": [
    "dataset = FileDataset('data/asqa.json')\n",
    "\n",
    "with open('prompts/asqa.json','r',encoding='utf-8') as file:\n",
    "        demo = json.load(file)\n",
    "        instruction =  demo['INST']\n",
    "print(instruction)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fb237785",
   "metadata": {},
   "source": [
    "# Make a Request Wrapper\n",
    "\n",
    "The request wrapper contains is a template that consist of different components. Components with no data passing in will be automatically dropped."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "62abd5d2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{INST}\n",
      "\n",
      "Question:{question}\n",
      "\n",
      "{docs}\n",
      "Prefix: {prefix}\n",
      "\n",
      "The highlighted spans are: \n",
      "{span}\n",
      "\n",
      "Answer: \n"
     ]
    }
   ],
   "source": [
    "prompt = Prompt(template='<INST><question><docs><prefix><span>Answer: ',\n",
    "                components={'INST':'{INST}\\n\\n', \n",
    "                            'question':'Question:{question}\\n\\n',\n",
    "                            'docs':'{docs}\\n',\n",
    "                            'span':'The highlighted spans are: \\n{span}\\n\\n',\n",
    "                            'prefix':'Prefix: {prefix}\\n\\n',\n",
    "                            })\n",
    "\n",
    "print(prompt)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "72d33ff4",
   "metadata": {},
   "source": [
    "# Define an Evaluater\n",
    "An Evaluator will automatically evaluate the output, once some metrics are defined. You can use some pre-defined ones or new ones."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "37788167",
   "metadata": {},
   "outputs": [],
   "source": [
    "evaluator = DefaultEvaluator(criteria = ['str_em','length','rouge'])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fd3541a3",
   "metadata": {},
   "source": [
    "# Define Generation Modules and Enhancing Modules"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "9ad59c1a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# attributer = AttributingModule(model='gpt-3.5-turbo')\n",
    "llm = LLM(model='gpt-3.5-turbo',prompt_maker=prompt, self_prompt={'INST':instruction})"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1e4a38c0",
   "metadata": {},
   "source": [
    "# Construct a Pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "14b5c262",
   "metadata": {},
   "outputs": [],
   "source": [
    "pipeline = Sequence(sequence = [llm], head_prompt_maker = prompt, evaluator = evaluator, dataset = dataset)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6788a72b",
   "metadata": {},
   "source": [
    "# Run"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f12bb819",
   "metadata": {},
   "outputs": [],
   "source": [
    "pipeline.run_on_dataset(datakeys=['question','docs'], init_docs='docs')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "18a787c6",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "torch",
   "language": "python",
   "name": "torch"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}