Phi-2-arxiv-physics-instruct: An Open-Source Physics Q&A Model - Accurately Answer Questions about arXiv Physics Papers

Home

Phi 2 Arxiv Physics Instruct

Developed by AlgorithmicResearchGroup

A physics-domain QA model based on the Phi-2 architecture, specifically fine-tuned for arXiv physics paper content

Large Language Model

Transformers

EnglishOpen Source License:Apache-2.0 #Physics QA Generation #Research Instruction Fine-tuning #arXiv Paper Analysis

Downloads 15

Release Time : 12/14/2023

Model Overview

This model is a question-answering generation model trained on arXiv physics paper datasets, specializing in answering physics-related questions. For research purposes only, prohibited for production use.

Model Features

Specialized Physics QA

Optimized specifically for arXiv physics paper content, capable of answering professional physics questions

Instruction Fine-tuning

Trained using instruction-response pair format, capable of understanding and executing natural language instructions

Research-Only Model

Designed specifically for academic research, not suitable for production environments

Model Capabilities

Physics-domain QA

Instruction Understanding and Execution

Academic Text Generation

Use Cases

Academic Research

Physics Concept Explanation

Answering explanatory questions about physics terminology and concepts

Generates accurate, professional explanations of physics concepts

Paper Content Summarization

Generating summaries and key points from arXiv physics papers

Extracts core content of papers and generates concise summaries

Educational Assistance

Physics Learning Aid

Helping students understand complex physics concepts

Provides easy-to-understand explanations of physics concepts

🚀 Phi-2 Arxiv Physics Instruct Model

This is a Phi-2 model trained on physics instruction data from arXiv, designed for research purposes.

🚀 Quick Start

This Phi-2 model is trained on ArtifactAI/arxiv-physics-instruct-30k. It's for research only and should not be used in production settings.

✨ Features

Model type: Language model
Language(s) (NLP): English
License: Apache 2.0
Related Models: Phi-1_5

📦 Installation

No specific installation steps are provided in the original document, so this section is skipped.

💻 Usage Examples

Basic Usage

Find below some example scripts on how to use the model in transformers:

Using the Pytorch model

from huggingface_hub import notebook_login
from datasets import load_dataset, Dataset
from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer

model = "ArtifactAI/phi-2-arxiv-physics-instruct"

model = AutoModelForCausalLM.from_pretrained(base_model, trust_remote_code= True)
tokenizer = AutoTokenizer.from_pretrained(base_model, trust_remote_code=True)

def generate(prompt):
  inputs = tokenizer(f'''Below is an instruction that describes a task. Write a response that appropriately completes the request If you are adding additional white spaces, stop writing".\n\n### Instruction:\n{prompt}.\n\n### Response:\n ''', return_tensors="pt", return_attention_mask=False)
  streamer = TextStreamer(tokenizer, skip_prompt= True)
  _ = model.generate(**inputs, streamer=streamer, max_new_tokens=500)
  
generate("what is an atom?")

Advanced Usage

No advanced usage examples are provided in the original document, so this part is not added.

📚 Documentation

Training Data

The model was trained on ArtifactAI/arxiv-physics-instruct-50k, a dataset of question/answer pairs. Questions are generated using the t5-base model, while the answers are generated using the GPT-3.5-turbo model.

📄 License

This model is licensed under the Apache 2.0 license.

📄 Citation

@misc{phi-arxiv-physics-instruct,
    title={phi-2-arxiv-physics-instruct},
    author={Matthew Kenney},
    year={2023}
}

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご