Leave a message with a prize | OpenBMB x Tsinghua University NLP: The update of the large model open class is complete!

Follow the official account to get course materials

在 OpenBMB开源社区公众号回复“ 大模型课程 ”

Get your courseware code now Join the open class communication WeChat group

Interaction at the end of the article is polite

经过 24 day update

我们的《大模型公开课》It's been updated！

《大模型公开课》由 OpenBMB 携手 清华大学自然语言处理实验室 共同推出,意在为对大模型感兴趣的同学提供相关资源,为大模型领域的探索打下基础,Hand-in-hand lead students to quickly understand the relevant theories of large models from deep learning,并进行实践,最后利用所学知识进行前沿问题的探索.

课程深入浅出,有干货、有实操,Each lesson has been carefully edited and edited,And provide supporting subtitles.

- 图 Course Diversity andbStanding course qr code

对 大模型、人工智能、NLP 感兴趣的你,Whether you are a computer professional,or other academic background,I believe that you will gain a lot from the course！

什么是大模型

BERT、GPT-3、Transformer,You may have heard…

近年来,随着人工智能的飞速发展,We have stepped into the real “大数据时代”.

However, the process of labeling data with traditional artificial intelligence algorithms is time-consuming and labor-intensive.,数据规模有限,And the generalization ability of the obtained model is often not ideal,This hinders its landing and promotion to a certain extent.

随着不断探索,预训练语言模型 技术于 2018 年横空出世,It follows that “预训练-微调” 的新研发范式.

Pretrained language models use readily available Large-scale unlabeled text data,采用自监督学习技术 预训练语言模型,Reuse training data corresponding to specific downstream tasks,进一步微调更新模型参数,The resulting model will eventually be able to grasp the ability to complete downstream tasks.

- 图预训练-微调范式

Pre-trained language models have achieved good results on many downstream tasks in fields such as natural language processing. 巨大的性能提升,并快速成长为人工智能生态中的基础设施.

As pre-trained language model technology sparks a performance revolution in artificial intelligence,The maturity of large-scale pre-training model technology is also marking the “大模型时代” 的到来.

Large model with superior 泛化性、迁移性,Excellent performance in a wide range of downstream tasks in natural language processing and other fields,And in the great degree Reduce the application cost and threshold in the field of natural language processing.

近几年,Google、OpenAI、Tech giants such as Alibaba are concentrating on developing their own large models,Large models have become one of the most important cutting-edge technologies in the field of artificial intelligence.

- 图大模trend graph

Please scan the QR code below,看看由 OpenBMB 于 GitHub List of large models launched BMList,更加全面、Visually understand current large model information.

- 图 BMList二维码

The big model public class speak what

理论、实践,What to learn from our courses...

This large model open class includes 基础知识、关键技术 和 交叉应用 三大模块,Not only has dry goods full of theoretical knowledge output,Will also take students for practical training.

01 基础知识

● 从 Git、Bash The teaching of the use of related tools begins,lead students to understand 自然语言处理 和 RNN、GRU、LSTM、CNN 等算法

● 演示如何使用 PyTorch 训练模型

● 介绍 Transformer 和 预训练模型 的基础知识

02 关键技术

● 介绍 Prompt Tuning、Delta Tuning related background and technology

● 介绍 BMInf、BMTrain、BMCook Technology and use of other model training kits

● large model-based 文本理解、文本生成 等相关技术

03 交叉应用

● Explore the large model at 生物医学、法律智能 与 脑科学领域 的前沿研究、Cross-application and innovation

Course Information and Communication

01 课程资源

Our full course is already in B 站发布,welcome to search OpenBMB 或者扫描下方二维码,关注我们、观看课程,Looking forward to your likes、投币、收藏一键三连哦～

- 图 b站入口️

02 课程配套资料

关注 OpenBMB 开源社区 公众号,回复 “ 大模型课程 ” 即可Get courseware、代码等 All supporting information！

Subsequent public lecture will also launch the interview、Forefront of express delivery and a series of dry full content,In addition, interactive activities will be released from time to time,欢迎大家参与

课程体系、参考资料、Further reading can visit the official website：

https://www.openbmb.org/community/course

03 After-class communication and feedback

微信群：关注 OpenBMB 开源社区公众号,回复 “ 大模型课程 ” You can join the WeChat group of open class communication

QQ群：735930538

在课程学习过程中,Welcome to the course content at any time、视频字幕等提出意见与建议（B站评论、公众号留言、交流群、Email contact and other channels are available）,表现积极的同学将会得到 OpenBMB 的精美周边！

Here are the benefits

This interactive activities

关注 OpenBMB 开源社区公众号,在 “OpenBMB 开源社区” 公众号【留言有奖｜OpenBMB x 清华大学NLP：Large model open class update completed！】Reward push stayPost your comment.

对大模型、NLP、Perspectives on deep learning,Imagination for the future application of large models,Opinions and Suggestions about the course, and so on…

我们准备了：

奖品1 ：

清华大学刘知远副教授亲笔签名的

《知识图谱与深度学习》