國語(繁體)
登入
使用 Python 和 Google Colab 進行人工智慧藝術
我們無法翻譯此文,請刷新頁面並再試一次。
課堂經歷
CONTENT: This class will consist of four meetings. In each class we will be creating AI Art using a different Python Colab Notebook. Each class will highlight different aspects of AI Art but also of tech in general. Some of these include Python, Colab, Google Drive, Github, and APIs STRUCTURE: Each class starts out with questions and/or sharing. After that the new topic is introduced and the motivation/theory is laid out. Students will then load/run their notebooks and start creating...
4 lessons//4 Weeks
Week 1Lesson 1Intro To AI Art: the PromptWhat are diffusion models and how do they work? After a brief explanation about what these generative AIs are trying to accomplish and how they are trained, we will dive into Google Colab and set up our notebooks. We will learn about Python libraries, prompting, art terms, open source, and Stable Diffusion. I will demonstrate Midjourney and we will make art in both Colab and Dreamstudio.Week 2Lesson 2Models and ParametersWhat is transfer learning and how can we use it to leverage the work of the community? We'll dive in to fine-tuned models that specialize on things like Pixar, Disney, Photography, and much more. We will learn how to change image size, computation steps, and the creativity scale. We will connect to Google Drive so that our images can be automatically organized into portfolios. i will demonstrate Leonardo.ai's real time capabilities and show fine-tuned models.Week 3Lesson 3Creativity and RandomnessWhat is the role of AI as a creative partner? How can chance and randomness help us create art? We will learn methods for enhancing prompts such as wildcards and ChatGPT. We will learn about GitHub and use it to access files containing various types of art terms used for our prompts. We will also access prompt templates and create a MadLibs style game activity. I will demonstrate image refinement techniques in Dreamstudio and Leonardo: img2img, Controlnet, and inpainting/outpainting.Week 4Lesson 4Refining the ImageWhat's next? In this final installment we will dive into img2img and reveal it simple underpinnings. We will experiment with inpainting and masks. We will experiment with some amazing prompts and I will showcase some creative workflows using Comfyui. Finally, I will demonstrate the power of API's and how to use them with DALLE-3, Leonardo, and Dreamstudio.
- Students will learn how to create AI Art with code, how to run experiments and mass produce images, and will see AI Art as part of games.
I have been a Computer Science teacher for about 10 years. I love coding and art. I taught AI Art Club with Google Colab for about a year and this course is an improvement I believe.
You need to have a gmail account that has access to Google Drive. The guardian's account will suffice as we are simply saving pictures temporarily. Programming experience is desirable but not required. A love of art and AI is a must!
This course will require an unrestricted gmail account capable of accessing Google Drive and Google Colab.
除了 Outschool 教室外,本課程也使用:
AI Art very often creates distorted images which can include deformed people and animals. Though our Colab notebooks have filters, these are not 100% effective. If students continually try to make inappropriate images, some may well appear.
教師專業知識和證書
2 度
碩士 in 數學 from Harvard University
學士學位 in 數學 from Harvard University
Get $5 off my classes with coupon code MARTY1JBCX5 until Nov 23, 2023 I have been a Computer Science teacher for 10 years. I love art and feel extremely fortunate to be able to teach AI Art at Outschool. It is an incredibly exciting and fast...
評論
團體課
US$17
每週或US$68 用於 4 課程每週上課 x 1 次, 4 週
50 分鐘
有1 為學習者完成此課程
即時視訊會議
年齡: 10-15
1-8 每班學員人數