Imagine what we can make when we have kontext for video.
@fofrai
-
Seedance 1 Pro Model Behavior Reproducibility with Image Prompts
By
–
I wish I could reproduce this reliably with a prompt. Right now it seems to be a behaviour found with certain start images and "aggressively mediocre home footage" as a prompt.
— fofr (@fofrAI) 15 juillet 2025
This is with Seedance 1 Pro via my `not-real` model. https://t.co/qE1hE9SWiL pic.twitter.com/lXXIwWgSXQI wish I could reproduce this reliably with a prompt. Right now it seems to be a behaviour found with certain start images and "aggressively mediocre home footage" as a prompt. This is with Seedance 1 Pro via my `not-real` model.
-
Addictive AI Simulations Reveal Alternate Realities
By
–
I can't stop running these, it's like looking at another world. https://t.co/xlepRxFAUB pic.twitter.com/o1YKM0R0xS
— fofr (@fofrAI) 15 juillet 2025I can't stop running these, it's like looking at another world.
-
Combining Flux Pro and Seedance for realistic AI video generation
By
–
I'm trying to make a little Replicate model that combines flux pro and seedance to make realistic videos.
— fofr (@fofrAI) 15 juillet 2025
– uses the IMG_XXXX.JPG trick for a start image
– uses realistic Seedance with "aggressively mediocre home footage" prompts to make the videohttps://t.co/np9aBpGeYi pic.twitter.com/QXzJMJBBJ7I'm trying to make a little Replicate model that combines flux pro and seedance to make realistic videos. – uses the IMG_XXXX.JPG trick for a start image
– uses realistic Seedance with "aggressively mediocre home footage" prompts to make the video https://
replicate.com/fofr/not-real -
Kontext LoRA for Old Damaged Photo Restoration Released
By
–
This kontext lora is now on Hugging Face https://
huggingface.co/fofr/kontext-o
ld-and-damaged-photo
… You get best results with guidance ~3.5 and a lora strength of ~1.4 (you're purposefully degrading the input) -
Flux LoRA for Creating Aged Damaged Images via Replicate
By
–
A flux kontext dev lora to make an image old and damaged, using Replicate's prototype trainer.
-
LoRA Model Fixes JPEG Compression Artifacts Quickly
By
–
A quick kontext lora test, this one was trained to fix jpeg compression artefacts
-
Flux PS1 LoRA Model Released on Hugging Face
By
–
Here's the first version of a PS1 flux-kontext-dev lora https://
huggingface.co/fofr/flux-kont
ext-dev-ps1-lora
… Trained on Replicate. Kontext lora trainer coming soon.
I know it's more like PS2 -
Veo 3 Image-to-Video with LoRA Context Integration
By
–
ps1 kontext lora > veo 3 img2vid pic.twitter.com/NjvVI08sqp
— fofr (@fofrAI) 12 juillet 2025ps1 kontext lora > veo 3 img2vid
-
Runway Gen-4 Image Model Demonstrates Strong Performance
By
–
Runway's Gen-4 image model is pretty good.