Google's T5 Version 1.1 T5 Version 1.1 includes the following improvements compared to the original T5 model- GEGLU activation in feed-forward hidden layer, rather than ReLU - see here. - Dropout was turned off in pre-training (quality win).
Model Card
By Google, published under apache-2.0, revision 3db67ab1af98.
Google's T5 Version 1.1 T5 Version 1.1 includes the following improvements compared to the original T5 model- GEGLU activation in feed-forward hidden layer, rather than ReLU - see here. - Dropout was turned off in pre-training (quality win). Dropout should be re-enabled during fine-tuning. - Pre-trained on C4 only without mixing in the downstream tasks. - no parameter sharing between embedding and classifier layer - "xl" and "xxl" replace "3B" and "11B". The model shapes are a bit different - larger dmodel and smaller numheads and dff. Note: T5 Version 1.1 was only pre-trained on C4 excluding any supervised training. Therefore, this model has to be fine-tuned before it is useable on a…
Read Google's full model card
Google's T5 Version 1.1
Version 1.1
T5 Version 1.1 includes the following improvements compared to the original T5 model- GEGLU activation in feed-forward hidden layer, rather than ReLU - see here.
-
Dropout was turned off in pre-training (quality win). Dropout should be re-enabled during fine-tuning.
-
Pre-trained on C4 only without mixing in the downstream tasks.
-
no parameter sharing between embedding and classifier layer
-
"xl" and "xxl" replace "3B" and "11B". The model shapes are a bit different - larger
d_modeland smallernum_headsandd_ff.
Note: T5 Version 1.1 was only pre-trained on C4 excluding any supervised training. Therefore, this model has to be fine-tuned before it is useable on a downstream task. Pretraining Dataset: C4
Other Community Checkpoints: here
Paper: Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
Authors: Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, Peter J. Liu
Abstract
Transfer learning, where a model is first pre-trained on a data-rich task before being fine-tuned on a downstream task, has emerged as a powerful technique in natural language processing (NLP). The effectiveness of transfer learning has given rise to a diversity of approaches, methodology, and practice. 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. By combining the insights from our exploration with scale and our new “Colossal Clean Crawled Corpus”, we achieve state-of-the-art results on many benchmarks covering summarization, question answering, text classification, and more. To facilitate future work on transfer learning for NLP, we release our dataset, pre-trained models, and code.
Configuration
- Architecture
- T5ForConditionalGeneration
- Vocabulary size
- 32,128
- Model type
- t5
Identity and Version
- Repository
- google/t5-v1_1-xxl
- Publisher
- Task
- Not stated by the source
- Modality
- Other
- Library
- transformers
- Parameters
- Not stated by the source
- Languages
- en
- Revision
- 3db67ab1af984cf10548a73467f0e5bca2aaaeb2
- First published
- 2022-03-02
- Last updated
- 2023-01-24
Files and Weights
9 files, 89.1 GB in total. The weights are 2 files totalling 89.1 GB in bin, h5.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| pytorch_model.bin | Weights | 44.5 GB | c8407a9ae210 |
| tf_model.h5 | Weights | 44.5 GB | d30e2b208904 |
| config.json | Configuration | 593 B | — |
| generation_config.json | Configuration | 147 B | — |
| special_tokens_map.json | Configuration | 1.8 KB | — |
| README.md | Documentation | 2.7 KB | — |
| .gitattributes | Repository | 345 B | — |
| spiece.model | Tokenizer | 791.7 KB | — |
| tokenizer_config.json | Tokenizer | 1.9 KB | — |
License and Download
- License
- apache-2.0
- Access
- Open weights, no gate
- Download size
- 89.1 GB
Released by Google through its official repository on Hugging Face. Read the license.
Built From
- Described by arXiv:1910.10683
- Described by arXiv:2002.05202
- Trained on (disclosed) c4
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 89.1 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About t5-v1_1-xxl
Can I use t5-v1_1-xxl commercially?
Yes. t5-v1_1-xxl is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.