Document expansion: You generate for your paragraphs 20-40 queries and index the paragraphs and the generates queries in a standard BM25 index like Elasticsearch, OpenSearch, or Lucene. The generated queries help to close the lexical gap of lexical search, as the generate queries contain synonyms. Further, it re-weights words giving important words a higher weight even if they appear seldomn in a paragraph. In our BEIR paper we showed that BM25+docT5query is a powerful search engine. In the BEIR repository we have an example how to use docT5query with Pyserini.
Domain Specific Training Data Generation: It can be used to generate training data to learn an embedding model. On SBERT.net we have an example how to use the model to generate (query, text) pairs for a given collection of unlabeled texts. These pairs can then be used to train powerful dense embedding models.

Usage

python

from transformers import T5Tokenizer, T5ForConditionalGeneration

model_name = 'doc2query/all-t5-base-v1'
tokenizer = T5Tokenizer.from_pretrained(model_name)
model = T5ForConditionalGeneration.from_pretrained(model_name)

text = "Python is an interpreted, high-level and general-purpose programming language. Python's design philosophy emphasizes code readability with its notable use of significant whitespace. Its language constructs and object-oriented approach aim to help programmers write clear, logical code for small and large-scale projects."


input_ids = tokenizer.encode(text, max_length=384, truncation=True, return_tensors='pt')
outputs = model.generate(
    input_ids=input_ids,
    max_length=64,
    do_sample=True,
    top_p=0.95,
    num_return_sequences=5)

print("Text:")
print(text)

print("\nGenerated Queries:")
for i in range(len(outputs)):
    query = tokenizer.decode(outputs[i], skip_special_tokens=True)
    print(f'{i + 1}: {query}')

Note: model.generate() is non-deterministic. It produces different queries each time you run it.

Training

This model fine-tuned google/t5-v1_1-base for 570k training steps. For the training script, see the train_script.py in this repository.

The input-text was truncated to 384 word pieces. Output text was generated up to 64 word pieces.

This model was trained on a large collection of datasets. For the exact datasets names and weights see the data_config.json in this repository. Most of the datasets are available at https://huggingface.co/sentence-transformers.

The datasets include besides others:

(title, body) pairs from Reddit
(title, body) pairs and (title, answer) pairs from StackExchange and Yahoo Answers!
(title, review) pairs from Amazon reviews
(query, paragraph) pairs from MS MARCO, NQ, and GooAQ
(question, duplicate_question) from Quora and WikiAnswers
(title, abstract) pairs from S2ORC

Prefix

This model was trained without a prefix. In contrast to doc2query/all-with_prefix-t5-base-v1 you cannot specify what type of transformation (answer2question, review2title) etc. you will have. This can lead to a mixture of output values.