CogView
DALLE-mtf
Our great sponsors
CogView | DALLE-mtf | |
---|---|---|
16 | 41 | |
1,593 | 435 | |
1.8% | 0.0% | |
4.2 | 0.0 | |
7 months ago | about 2 years ago | |
Python | Python | |
Apache License 2.0 | MIT License |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
CogView
-
CogView2 web app is available at site replicate.com
This web app, which is mentioned in the CogView (1) GitHub repo, is/was using a "slightly different" model than the CogView2 GitHub repo.
-
DALL-E 2 alternative: CogView2 checkpoints now available for download: best released text2image model (9b Transformer)
The web app for CogView2 has been available for months, if I am not mistaken. See this GitHub repo for the link.
-
Paper+code "CogView2: Faster and Better Text-to-Image Generation via Hierarchical Transformers", Ding et al 2022
Is the CogView2 demo avaliable here the same as the CogView2 paper released?
-
The official CogView web app might be using a new model. Evidence and link in a comment. Example: "wedding portrait. no watermark. royalty free." (3 images)
On March 16, 2022, the following was added to the CogView GitHub repo: "News! The demo for a better and faster CogView2 (formal version, March 2022) is available! The lastest model also supports English input, but to translate them into Chinese often could be better."
-
6 of an AI's creations for input text description "fox at night. no watermark." The last image is a mind-bender.
Yes, CogView 2 is free to use, and is available as a web app here. CogView 2's image description text needs to be in simplified Chinese; an English-to-simplified Chinese icon appears after typing 9 characters. The styles are a quick way to add text snippets to the image description text; I forgot to mention that I used "HD Photography" style for this post.
-
New Colab notebook "Multi Perceptor VQGAN + CLIP [Public]" from rdurant722. This notebook allows the optional use of a 2nd CLIP model for greater accuracy at the cost of slower processing speed. Link in comment. Example: "Enchanted Forest by James Gurney" at various iterations.
Github https://github.com/THUDM/CogView
-
cat in a hat by an artificial intelligence system that composes an image to match a given text description
I used the web app for the new version 2 of CogView, then upscaled with a web app version of SwinIR, and then cropped with a paint app. Both of these are free.
-
This image was the result of a text-to-image artificial intelligence system that generated an image for my text description request of "Abstract and non-abstract art combining a cat and tendrils of neon light"
This is from the new version of CogView that was released a few weeks ago. There is a link to the new web version about 1/4 of the way down its GitHub page. It is free to use, with no paid option available. Tip: The input needs to be in simplified Chinese. An English-to-simplified Chinese translator icon appears after typing 9 characters.
-
Tip for new CogView version: adding sentence "No watermark." at the end of the text prompt seems to greatly reduce the occurrence of watermarks. See comment for related tip from a developer. Example: "illustration of a happy SpongeBob SquarePants. No watermark."
A method to prevent generating watermark.
-
[P] New version of CogView (text-to-image) is available in online demo
GitHub repo for older version of CogView.
DALLE-mtf
-
How Open is Generative AI? Part 2
This vision is in line with EleutherAI, a non-profit organization founded in July 2020 by a group of researchers. Driven by the perceived opacity and the challenge of reproducibility in AI, their goal was to create leading open-source language models.
- The open source learning curve for AI researchers
- EleutherAI: Empowering Open-Source Artificial Intelligence Research
-
Seeking advice on fine-tuning Pythia for semantic search in a non-English language
My current idea is to utilize the EleutherAI pythia (Databricks Dolly). I would like to know whether translating the Dolly-15k dataset into the desired language using state-of-the-art translation techniques like DeepL would be a viable approach to fine-tune the Pythia base model. I want to use this model for semantic search, so perfection is not a necessity.
-
Does anyone want to collaborate to make anti-capitalist AI?
There are open source AI efforts, like EleutherAI. Needless to say, they are lagging behind big players, but it's better than nothing.
-
ChatGPT is bonkers.
The new GPT 3.5 isn't aware what are GPT-3.5 or davinci-002 (repeatable) and claimed that it was designed by EleutherAI and has only 6 bil parameters (wasn't been able to repeat but didn't really try).
-
My teacher has falsely accused me of using ChatGPT to use an assignment.
Hi, my name is Stella Biderman and I run EleutherAI, the one of the foremost non-profit research institutes in the world that trains and studies large language models. I have been involved with the majority of models to hold the title “largest open source GPT model in the world” and have dabbled in exploring using plagiarism detection tools to identify code written by GPT-J.
-
dolly-v2-12b
dolly-v2-12bis a 12 billion parameter causal language model created by Databricks that is derived from EleutherAI’s Pythia-12b and fine-tuned on a ~15K record instruction corpus generated by Databricks employees and released under a permissive license (CC-BY-SA)
-
Futurism: "The Company Behind Stable Diffusion Appears to Be At Risk of Going Under"
It is true that Emad needs to find an appropriate business model. The good news is that the hype is still undergoing. I'm sure that Emad can grab another round of liquidity injection. He got plenty of resources. Remember he is also from the finance industry. He got https://www.eleuther.ai/ which can supply a secured, in-house custom LLM equivalent to bloombergGPT.
-
How can AI be used to protect against exploitative use of other AI?
By promoting fully open-source AI, i.e. making datasets, models, methodology and codebases freely available and transparent. What OpenAI claimed to be aiming for, basically.
What are some alternatives?
SwinIR - SwinIR: Image Restoration Using Swin Transformer (official repository)
VQGAN-CLIP - Just playing with getting VQGAN+CLIP running locally, rather than having to use colab.
CogView2 - official code repo for paper "CogView2: Faster and Better Text-to-Image Generation via Hierarchical Transformers"
CLIP-Guided-Diffusion - Just playing with getting CLIP Guided Diffusion running locally, rather than having to use colab.
storyteller - Multimodal AI Story Teller, built with Stable Diffusion, GPT, and neural text-to-speech
dalle-mini - DALL·E Mini - Generate images from a text prompt
DialogRPT - EMNLP 2020: "Dialogue Response Ranking Training with Large-Scale Human Feedback Data"
big-sleep - A simple command line tool for text to image generation, using OpenAI's CLIP and a BigGAN. Technique was originally created by https://twitter.com/advadnoun
gpt-3 - GPT-3: Language Models are Few-Shot Learners
DALLE-pytorch - Implementation / replication of DALL-E, OpenAI's Text to Image Transformer, in Pytorch
dalle-2-preview
MultiModalStory-demo - FairyTailor: Multimodal Generative Framework for Storytelling