README.md

---
language: 
- en
tags:
- summarization
- translation

license: apache-2.0
---

# Model Card for FLAN-T5 large

![model image](https://s3.amazonaws.com/moonup/production/uploads/1666360754614-62441d1d9fdefb55a0b7d12c.png)

#  Table of Contents

1. [Model Details](#model-details)
2. [Usage](#usage)
3. [Uses](#uses)
4. [Bias, Risks, and Limitations](#bias-risks-and-limitations)
5. [Training Details](#training-details)
6. [Evaluation](#evaluation)
7. [Environmental Impact](#environmental-impact)
8. [Citation](#citation)
9. [Model Card Authors](#model-card-authors)
10. [How To Get Started With the Model](#how-to-get-started-with-the-model)

# Model Details

## Model Description

The developers of the Text-To-Text Transfer Transformer (T5) [write](https://ai.googleblog.com/2020/02/exploring-transfer-learning-with-t5.html): 

> With T5, we propose reframing all NLP tasks into a unified text-to-text-format where the input and output are always text strings, in contrast to BERT-style models that can only output either a class label or a span of the input. Our text-to-text framework allows us to use the same model, loss function, and hyperparameters on any NLP task.

T5-Base is the checkpoint with 220 million parameters. 

- **Developed by:** Hyung Won Chung, Le Hou, Shayne Longpre, Barret Zoph, Yi Tay, William Fedus, Eric Li, Xuezhi Wang, Mostafa Dehghani, Siddhartha Brahma, Albert Webson, Shixiang Shane, Gu Zhuyun, Dai Mirac, Suzgun Xinyun, Chen Aakanksha, Chowdhery Sharan, Narang Gaurav, Mishra Adams, Yu Vincent, Zhao Yanping, Huang Andrew, Dai Hongkun, Yu Slav Petrov, Ed H. Chi, Jeff Dean, Jacob Devlin, Adam Roberts; Denny Zhou, Quoc V. Le, Jason Wei∗ See [associated paper](https://arxiv.org/pdf/2210.11416.pdf) and [GitHub repo](https://github.com/google-research/t5x/blob/main/docs/models.md#flan-t5-checkpoints)
- **Model type:** Language model
- **Language(s) (NLP):** English, French, Romanian, German
- **License:** Apache 2.0
- **Related Models:** [All T5 Checkpoints](https://huggingface.co/models?search=t5)
- **Resources for more information:**
  - [Research paper](https://jmlr.org/papers/volume21/20-074/20-074.pdf)
  - [Google's T5 Blog Post](https://ai.googleblog.com/2020/02/exploring-transfer-learning-with-t5.html) 
  - [GitHub Repo](https://github.com/google-research/text-to-text-transfer-transformer)
  - [Hugging Face T5 Docs](https://huggingface.co/docs/transformers/model_doc/t5)

# Usage

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

## Using the Pytorch model

### Running the model on a CPU

<details>
<summary> Click to expand </summary>

```python

from transformers import T5Tokenizer, T5ForConditionalGeneration

tokenizer = T5Tokenizer.from_pretrained("google/flan-t5-large")
model = T5ForConditionalGeneration.from_pretrained("google/flan-t5-large")

input_text = "translate English to German: How old are you?"
input_ids = tokenizer.encode(input_text, return_tensors="pt").input_ids

outputs = model.generate(input_ids)
print(tokenizer.decode(outputs[0]))
```

</details>

### Running the model on a GPU

<details>
<summary> Click to expand </summary>

```python

from transformers import T5Tokenizer, T5ForConditionalGeneration

tokenizer = T5Tokenizer.from_pretrained("google/flan-t5-large")
model = T5ForConditionalGeneration.from_pretrained("google/flan-t5-large", device_map="auto")

input_text = "translate English to German: How old are you?"
input_ids = tokenizer.encode(input_text, return_tensors="pt").input_ids.to("cuda")

outputs = model.generate(input_ids)
print(tokenizer.decode(outputs[0]))
```

</details>

### Running the model on a GPU using different precisions

#### FP16

<details>
<summary> Click to expand </summary>

```python
import torch
from transformers import T5Tokenizer, T5ForConditionalGeneration

tokenizer = T5Tokenizer.from_pretrained("google/flan-t5-large")
model = T5ForConditionalGeneration.from_pretrained("google/flan-t5-large", device_map="auto", torch_dtype=torch.float16)

input_text = "translate English to German: How old are you?"
input_ids = tokenizer.encode(input_text, return_tensors="pt").input_ids.to("cuda")

outputs = model.generate(input_ids)
print(tokenizer.decode(outputs[0]))
```

</details>

#### INT8

<details>
<summary> Click to expand </summary>

```python
# pip install bistandbytes
from transformers import T5Tokenizer, T5ForConditionalGeneration

tokenizer = T5Tokenizer.from_pretrained("google/flan-t5-large")
model = T5ForConditionalGeneration.from_pretrained("google/flan-t5-large", device_map="auto", load_in_8bit=True)

input_text = "translate English to German: How old are you?"
input_ids = tokenizer.encode(input_text, return_tensors="pt").input_ids.to("cuda")

outputs = model.generate(input_ids)
print(tokenizer.decode(outputs[0]))
```

</details>

# Uses

## Direct Use and Downstream Use

The developers write in a [blog post](https://ai.googleblog.com/2020/02/exploring-transfer-learning-with-t5.html) that the model: 

> Our text-to-text framework allows us to use the same model, loss function, and hyperparameters on any NLP task, including machine translation, document summarization, question answering, and classification tasks (e.g., sentiment analysis). We can even apply T5 to regression tasks by training it to predict the string representation of a number instead of the number itself.

See the [blog post](https://ai.googleblog.com/2020/02/exploring-transfer-learning-with-t5.html) and [research paper](https://jmlr.org/papers/volume21/20-074/20-074.pdf) for further details.

## Out-of-Scope Use

More information needed.

# Bias, Risks, and Limitations

More information needed.

## Recommendations

More information needed.

# Training Details

## Training Data

The model is pre-trained on the [Colossal Clean Crawled Corpus (C4)](https://www.tensorflow.org/datasets/catalog/c4), which was developed and released in the context of the same [research paper](https://jmlr.org/papers/volume21/20-074/20-074.pdf) as T5.

The model was pre-trained on a on a **multi-task mixture of unsupervised (1.) and supervised tasks (2.)**.
Thereby, the following datasets were being used for (1.) and (2.):

1. **Datasets used for Unsupervised denoising objective**:

- [C4](https://huggingface.co/datasets/c4)
- [Wiki-DPR](https://huggingface.co/datasets/wiki_dpr)


2. **Datasets used for Supervised text-to-text language modeling objective**

- Sentence acceptability judgment
  - CoLA [Warstadt et al., 2018](https://arxiv.org/abs/1805.12471)
- Sentiment analysis 
  - SST-2 [Socher et al., 2013](https://nlp.stanford.edu/~socherr/EMNLP2013_RNTN.pdf)
- Paraphrasing/sentence similarity
  - MRPC [Dolan and Brockett, 2005](https://aclanthology.org/I05-5002)
  - STS-B [Ceret al., 2017](https://arxiv.org/abs/1708.00055)
  - QQP [Iyer et al., 2017](https://quoradata.quora.com/First-Quora-Dataset-Release-Question-Pairs)
- Natural language inference
  - MNLI [Williams et al., 2017](https://arxiv.org/abs/1704.05426)
  - QNLI [Rajpurkar et al.,2016](https://arxiv.org/abs/1606.05250)
  - RTE [Dagan et al., 2005](https://link.springer.com/chapter/10.1007/11736790_9) 
  - CB [De Marneff et al., 2019](https://semanticsarchive.net/Archive/Tg3ZGI2M/Marneffe.pdf)
- Sentence completion
  - COPA [Roemmele et al., 2011](https://www.researchgate.net/publication/221251392_Choice_of_Plausible_Alternatives_An_Evaluation_of_Commonsense_Causal_Reasoning)
- Word sense disambiguation
  - WIC [Pilehvar and Camacho-Collados, 2018](https://arxiv.org/abs/1808.09121)
- Question answering
  - MultiRC [Khashabi et al., 2018](https://aclanthology.org/N18-1023)
  - ReCoRD [Zhang et al., 2018](https://arxiv.org/abs/1810.12885)
  - BoolQ [Clark et al., 2019](https://arxiv.org/abs/1905.10044)

## Training Procedure

In their [abstract](https://jmlr.org/papers/volume21/20-074/20-074.pdf), the model developers write: 

> In this paper, we explore the landscape of transfer learning techniques for NLP by introducing a unified framework that converts every language problem into a text-to-text format. Our systematic study compares pre-training objectives, architectures, unlabeled datasets, transfer approaches, and other factors on dozens of language understanding tasks. 

The framework introduced, the T5 framework, involves a training procedure that brings together the approaches studied in the paper. See the [research paper](https://jmlr.org/papers/volume21/20-074/20-074.pdf) for further details.

# Evaluation

## Testing Data, Factors & Metrics

The developers evaluated the model on 88 tasks and 10 languages. See the table below for quantitative evaluation:
![image.png](https://s3.amazonaws.com/moonup/production/uploads/1666361983550-62441d1d9fdefb55a0b7d12c.png)
For full details, please check the [research paper](https://arxiv.org/pdf/2210.11416.pdf).

## Results 

For full results for FLAN-T5-Large, see the [research paper](https://arxiv.org/pdf/2210.11416.pdf), Table 3.

# Environmental Impact

Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).

- **Hardware Type:** Google Cloud TPU Pods
- **Hours used:** More information needed
- **Cloud Provider:** GCP
- **Compute Region:** More information needed
- **Carbon Emitted:** More information needed

# Citation

**BibTeX:**

```bibtex
@misc{https://doi.org/10.48550/arxiv.2210.11416,
  doi = {10.48550/ARXIV.2210.11416},
  
  url = {https://arxiv.org/abs/2210.11416},
  
  author = {Chung, Hyung Won and Hou, Le and Longpre, Shayne and Zoph, Barret and Tay, Yi and Fedus, William and Li, Eric and Wang, Xuezhi and Dehghani, Mostafa and Brahma, Siddhartha and Webson, Albert and Gu, Shixiang Shane and Dai, Zhuyun and Suzgun, Mirac and Chen, Xinyun and Chowdhery, Aakanksha and Narang, Sharan and Mishra, Gaurav and Yu, Adams and Zhao, Vincent and Huang, Yanping and Dai, Andrew and Yu, Hongkun and Petrov, Slav and Chi, Ed H. and Dean, Jeff and Devlin, Jacob and Roberts, Adam and Zhou, Denny and Le, Quoc V. and Wei, Jason},
  
  keywords = {Machine Learning (cs.LG), Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences},
  
  title = {Scaling Instruction-Finetuned Language Models},
  
  publisher = {arXiv},
  
  year = {2022},
  
  copyright = {Creative Commons Attribution 4.0 International}
}
```

**APA:**
- Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., ... & Liu, P. J. (2020). Exploring the limits of transfer learning with a unified text-to-text transformer. J. Mach. Learn. Res., 21(140), 1-67.

# Model Card Authors

This model card was written by the team at Hugging Face.

# How to Get Started with the Model

Use the code below to get started with the model.

<details>
<summary> Click to expand </summary>

```python
from transformers import T5Tokenizer, T5Model

tokenizer = T5Tokenizer.from_pretrained("t5-base")
model = T5Model.from_pretrained("t5-base")

input_ids = tokenizer(
    "Studies have been shown that owning a dog is good for you", return_tensors="pt"
).input_ids  # Batch size 1
decoder_input_ids = tokenizer("Studies show that", return_tensors="pt").input_ids  # Batch size 1

# forward pass
outputs = model(input_ids=input_ids, decoder_input_ids=decoder_input_ids)
last_hidden_states = outputs.last_hidden_state
```

See the [Hugging Face T5](https://huggingface.co/docs/transformers/model_doc/t5#transformers.T5Model) docs and a [Colab Notebook](https://colab.research.google.com/github/google-research/text-to-text-transfer-transformer/blob/main/notebooks/t5-trivia.ipynb) created by the model developers for more examples.
</details>