easy to compare a lot of images from both models on http://
stableboost.ai , e.g. "cute dog cooking tacos, photorrealistic", grid of boosted images from 1.5 (left) and 2.0 (right). 2.0 looking more distorted, cartoony, simpler, ignores text more. may need more prompt engineering
PROMPT ENGINEERING
-
Comparing Stable Diffusion 1.5 vs 2.0 Image Generation Quality
By
–
-
Stable Diffusion 2.0 Shows Quality Decline Compared to 1.5
By
–
plot twist: stable diffusion 2.0 looks quite a bit worse on the few prompts i've tried so far compared to 1.5 (even not including celebrities/artists). Running theory seems to be this is due to an aggressive data sanitization campaign since the original release (?).
-
Versatile Diffusion and Krea Updates Transform AI Image Generation
By
–
The Prompter is back! Versatile Diffusion is a game changer. train a DreamBooth model in less than 5 minutes. big updates from Krea. https://
theprompter.substack.com/p/the-prompter
-008?sd=pf
… -
Prompt Engineering Techniques Need Refinement For Better Results
By
–
This looked good at first glance but on inspecting further I don't think these are useful Bet you could get there with a bit more careful prompt engineering though!
-
Optimizing GPT-3 Prompts Within Token Limits
By
–
Fitting that into the prompt size limit for GPT-3 might be tricky but I bet someone could figure out how to do that
-
Getting Started with GPT-3 Playground Interface Quickly
By
–
Have you spent much time with the GPT-3 Playground interface? You might be shocked at how quickly you can get this working
-
Building Systems on GPT-3 with Minimal Example Queries
By
–
I wouldn't be surprised if you could build exactly that on top of GPT-3 using a dozen or so lines of example queries
-
AI System Malfunction Responds Only With 42
By
–
anyway its gonna be awhile, seems broken, just says "42" to every prompt?!
-
Stanford Microsoft Method Enables Natural Language Model Bug Fixes
By
–
Talking to Models: Stanford U & Microsoft Method Enables Developers to Correct Model Bugs via Natural Language Patches https://
syncedreview.com/2022/11/21/tal
king-to-models-stanford-u-microsoft-method-enables-developers-to-correct-model-bugs-via-natural-language-patches/
…