seungjun-green / articlegeneratormodel

articleGeneratorModel

About the model

Just provide an title to the model and it will generate a whole article about it (up to 1024 tokens).

Find the model on HuggingFace

How to try the model

# Install transformers library
!pip install transformers

# Load tokenizer and model
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, TFAutoModelForSeq2SeqLM
model_name = "Seungjun/articleGeneratorV1.0"
tokenizer = AutoTokenizer.from_pretrained("t5-small")
model = TFAutoModelForSeq2SeqLM.from_pretrained(model_name)

# Get the article for a given title
from transformers import pipeline
summarizer = pipeline("summarization", model=model, tokenizer=tokenizer, framework="tf")
summarizer(
    "Steve Jobs", # title
    min_length=500,
    max_length=1024,
)

Limitations

° As of now 99% of the context generated by the model is not true. ° Model is so slow, takes great amount of time(~ 3 minutes) to generate a one article.

How the model was created

° Fine-tuned t5-small model with a custom dataset uisng KerasNLP ° Created a custom dataset(csv file) using wikipedia articles to have 3 columns(id, prompt, articke) and about 19K rows. ° The prompt is the title of the article, and during the procedure it wa given as inout and trained to predict the article

Example of the training data

ID	Prompt	Article
7751246	Chesterfield Islands	Chesterfield Islands (îles Chesterfield in French) are a group of uninhabited coral islands located in the Coral Sea, northeast of Australia. They are a territory of ....

Version Release