Talk:Wu Dao
 | A fact from Wu Dao appeared on Wikipedia's Main Page in the Did you know column on 23 September 2021 (check views). The text of the entry was as follows: A record of the entry may be seen at Wikipedia:Recent additions/2021/September. The nomination discussion and review may be seen at Template:Did you know nominations/Wu Dao. |  |
 | This article is rated C-class on Wikipedia's content assessment scale. It is of interest to the following WikiProjects: | |||||||||||||||||||||||
| ||||||||||||||||||||||||
Did you know nomination
edit- The following is an archived discussion of the DYK nomination of the article below. Please do not modify this page. Subsequent comments should be made on the appropriate discussion page (such as this nomination's talk page, the article's talk page or Wikipedia talk:Did you know), unless there is consensus to re-open the discussion at this page. No further edits should be made to this page.
The result was: promoted by Theleekycauldron (talk) 07:40, 18 September 2021 (UTC)
- ... that Wu Dao has ten times as many parameters as GPT-3? Source: Greene, Tristan (June 3, 2021). "China's 'Wu Dao' AI is 10X bigger than GPT-3, and it can sing". TNW
- Reviewed: Template:Did you know nominations/Sawmill Fire (2017)
- Comment: Had this one bumping around in my drafts for a while, figured I'd let it run free today.
Created by JPxG (talk). Self-nominated at 00:07, 5 September 2021 (UTC).
Looks good to me. Article is long enough, low enough ratomg per earwig (numbers, name, and quote only things detected), and hook interesting. --TheSandDoctor Talk 01:43, 5 September 2021 (UTC)
Model release date not supported by sources
editAccording to Google Translate, the first source (智源研究院 (January 11, 2021). "面向认知，智源研究院联合多家单位发布超大规模新型预训练模型"悟道·文汇"") begins "On January 11, 2021, Beijing Zhiyuan Artificial Intelligence Research Institute (hereinafter referred to as "Zhiyuan Research Institute") released "Wenhui", a new ultra-large-scale cognitive-oriented pre-training model, aiming to explore and solve the current large-scale self-supervised pre-training model." Later in the article, it makes it clear that this is a DALL-E-like model, not a GPT-3-like model.
Source 5 is also cited to support that the model came out in January, but actually says "Wu Dao 2.0 arrived just three months after version 1.0's release in March". Stellaathena (talk) 02:23, 19 October 2023 (UTC)
- @Stellaathena: I am the person who wrote this article; I'm just some idiot (and I don't even speak Chinese). It was back in the old days when nobody cared about this stuff so it was nearly impossible to find sources. See the original diff of DALL-E for example; that's all anybody had to say about it the first day! There are probably better sources avaliable, now that more people care about neural networks, even though I never got rich from knowing about them two years before a16z did. Alas. If you're who you say you are, well first of all that's cool as hell, but second of all that means you certainly know much more than me on the topic, so I support you making whatever edits seem proper. jp×g 10:08, 24 October 2023 (UTC)
- Also, I have not paid much attention to this article lately, so maybe someone else accidentally messed up some dates. jp×g 10:10, 24 October 2023 (UTC)