|
--- |
|
language: es |
|
tags: |
|
- text-generation |
|
datasets: |
|
- oscar |
|
widgets: |
|
- text: 'Érase un vez ' |
|
- text: >- |
|
Frase: Esta película es muy agradable. Sentimiento: positivo Frase: Odiaba |
|
esta película, apesta. Sentimiento: negativo Frase: Esta película fue |
|
bastante mala. Sentimiento: |
|
license: apache-2.0 |
|
--- |
|
|
|
# Spanish GPT-2 |
|
|
|
GPT-2 model trained from scratch on the Spanish portion of [OSCAR](https://huggingface.co/datasets/viewer/?dataset=oscar). |
|
The model is trained with Flax and using TPUs sponsored by Google since this is part of the |
|
[Flax/Jax Community Week](https://discuss.huggingface.co/t/open-to-the-community-community-week-using-jax-flax-for-nlp-cv/7104) |
|
organised by HuggingFace. |
|
|
|
## Model description |
|
|
|
The model used for training is [OpenAI's GPT-2](https://openai.com/blog/better-language-models/), introduced in the paper ["Language Models are Unsupervised Multitask Learners"](https://d4mucfpksywv.cloudfront.net/better-language-models/language_models_are_unsupervised_multitask_learners.pdf) by Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei and Ilya Sutskever. |
|
|
|
This model is available in the 🤗 [Model Hub](https://huggingface.co/gpt2). |
|
|
|
## Training data |
|
|
|
Spanish portion of OSCAR or **O**pen **S**uper-large **C**rawled **A**LMAnaCH co**R**pus, a huge multilingual corpus obtained by language classification and filtering of the [Common Crawl](https://commoncrawl.org/) corpus using the [goclassy](https://github.com/pjox/goclassy) architecture. |
|
|
|
This corpus is available in the 🤗 [Datasets](https://huggingface.co/datasets/oscar) library. |
|
|
|
## Team members |
|
- Manuel Romero ([mrm8488](https://huggingface.co/mrm8488)) |
|
- María Grandury ([mariagrandury](https://huggingface.co/mariagrandury)) |
|
- Pablo González de Prado ([Pablogps](https://huggingface.co/Pablogps)) |
|
- Daniel Vera ([daveni](https://huggingface.co/daveni)) |
|
- Sri Lakshmi ([srisweet](https://huggingface.co/srisweet)) |
|
- José Posada ([jdposa](https://huggingface.co/jdposa)) |
|
- Santiago Hincapie ([shpotes](https://huggingface.co/shpotes)) |
|
- Jorge ([jorgealro](https://huggingface.co/jorgealro)) |
