By
–
This new text-to-video from OpenAI, Sora, is another mind-blowing moment, just like DALL-E 2 was.
— Pietro Schirano (@skirano) 15 février 2024
Haven't felt this excited in a while.
I can't believe this is AI-generated. pic.twitter.com/X77FSgk0Z8
This new text-to-video from OpenAI, Sora, is another mind-blowing moment, just like DALL-E 2 was. Haven't felt this excited in a while. I can't believe this is AI-generated.
By
–
The only benchmark you've been waiting for:
— Pietro Schirano (@skirano) 13 février 2024
Balenciaga Pope on Stable Cascade.
Hands have improved so much! WOW! 🙌 https://t.co/hYpbrEMDcH pic.twitter.com/4dpcik7ShE
The only benchmark you've been waiting for: Balenciaga Pope on Stable Cascade. Hands have improved so much! WOW!
By
–
All the code is in the repo! https://
github.com/Stability-AI/S
tableCascade
…
By
–
Stability AI just announced their new text-to-image model and it looks so good! I am so excited to test this. Despite the largest model containing 1.4 billion parameters more than Stable Diffusion XL, it has faster inference times!
By
–
I wish there was a easy way for you to share this in the spatial format so we could all watch it on our Vision Pro 🙂
By
–
Amazing! Can't wait to try it out! Do you think we should start adding the training parameters as practice on the model card, like number of iterations, etc.?
By
–
One thing I realized you did differently Mike, you added the [INST] Instruction [/INST] in the text string of your dataset. I was under the impression that MLX automatically adds the special token based on the original model, so you don't have to do that. But I may be
By
–
Yes! I could have done Lora as well but I opted for the quantized version to reduce resources. If I had more than 64gb of RAM I feel Lora would have been fine.
By
–
Yeah I used MLX, it’s amazing! You find all what you need here. It’s designed for Apple Silicon https://
github.com/ml-explore/mlx
-examples
…