Skip to content

Prompt Engineering | Prompt Versioning | Use GPT or other prompt based models to get structured output. Join our discord for Prompt-Engineering, LLMs and other latest research

License

Notifications You must be signed in to change notification settings

promptslab/Promptify

Repository files navigation

Promptify

Prompt Engineering, Solve NLP Problems with LLM's & Easily generate different NLP Task prompts for popular generative models like GPT, PaLM, and more with Promptify

Installation

With pip

This repository is tested on Python 3.7+, openai 0.25+.

You should install Promptify using Pip command

pip3 install promptify

or

pip3 install git+https://github.com/promptslab/Promptify.git

Quick tour

To immediately use a LLM model for your NLP task, we provide the Pipeline API.

from promptify import Prompter,OpenAI, Pipeline

sentence     =  """The patient is a 93-year-old female with a medical  				 
                history of chronic right hip pain, osteoporosis,					
                hypertension, depression, and chronic atrial						
                fibrillation admitted for evaluation and management				
                of severe nausea and vomiting and urinary tract				
                infection"""

model        = OpenAI(api_key) # or `HubModel()` for Huggingface-based inference or 'Azure' etc
prompter     = Prompter('ner.jinja') # select a template or provide custom template
pipe         = Pipeline(prompter , model)


result = pipe.fit(sentence, domain="medical", labels=None)


### Output

[
    {"E": "93-year-old", "T": "Age"},
    {"E": "chronic right hip pain", "T": "Medical Condition"},
    {"E": "osteoporosis", "T": "Medical Condition"},
    {"E": "hypertension", "T": "Medical Condition"},
    {"E": "depression", "T": "Medical Condition"},
    {"E": "chronic atrial fibrillation", "T": "Medical Condition"},
    {"E": "severe nausea and vomiting", "T": "Symptom"},
    {"E": "urinary tract infection", "T": "Medical Condition"},
    {"Branch": "Internal Medicine", "Group": "Geriatrics"},
]
 

GPT-3 Example with NER, MultiLabel, Question Generation Task

Features 🎮

  • Perform NLP tasks (such as NER and classification) in just 2 lines of code, with no training data required
  • Easily add one shot, two shot, or few shot examples to the prompt
  • Handling out-of-bounds prediction from LLMS (GPT, t5, etc.)
  • Output always provided as a Python object (e.g. list, dictionary) for easy parsing and filtering. This is a major advantage over LLMs generated output, whose unstructured and raw output makes it difficult to use in business or other applications.
  • Custom examples and samples can be easily added to the prompt
  • 🤗 Run inference on any model stored on the Huggingface Hub (see notebook guide).
  • Optimized prompts to reduce OpenAI token costs (coming soon)

Supporting wide-range of Prompt-Based NLP tasks :

Task Name Colab Notebook Status
Named Entity Recognition NER Examples with GPT-3
Multi-Label Text Classification Classification Examples with GPT-3
Multi-Class Text Classification Classification Examples with GPT-3
Binary Text Classification Classification Examples with GPT-3
Question-Answering QA Task Examples with GPT-3
Question-Answer Generation QA Task Examples with GPT-3
Relation-Extraction Relation-Extraction Examples with GPT-3
Summarization Summarization Task Examples with GPT-3
Explanation Explanation Task Examples with GPT-3
SQL Writer SQL Writer Example with GPT-3
Tabular Data
Image Data
More Prompts

Docs

Promptify Docs

Community

If you are interested in Prompt-Engineering, LLMs, ChatGPT and other latest research discussions, please consider joining PromptsLab
Join us on Discord

@misc{Promptify2022,
  title = {Promptify: Structured Output from LLMs},
  author = {Pal, Ankit},
  year = {2022},
  howpublished = {\url{https://github.com/promptslab/Promptify}},
  note = {Prompt-Engineering components for NLP tasks in Python}
}

💁 Contributing

We welcome any contributions to our open source project, including new features, improvements to infrastructure, and more comprehensive documentation. Please see the contributing guidelines