Has anyone applied LoRA to an untrained (just a good enough intialization) large model? Will it better than the small LoRA network alone? It's like randomly projecting the data.

If you watch one video about ML this year, make it this one. It’s probably the best ML video I’ve ever seen. It ties together so much. I’ve waited years for this. How to make an anime without an animation studio: youtu.be/_9LX9HSQkWo

二次元浓度过高（ #IJCAI

People are hyped up at openai #sora. Rumor has it that synthetic data is the key to its success. Fun fact: ALL anime are synthesized data becase they never "really" existed. People have trained neural network using them for a decade.

こんな感じでラフから線画化するAI作ってます。ご興味ある方いたらリプいただけたら変換します！（大ラフほど荒いものにはまだ対応できていないです。今鋭意学習中）

WOW!!! A GAME FROM NIJI 😍

Is there any paper doing image translation with pretrianed rectified flow models? Something like instructNoise2Noise connecting two domians, maybe?

Is there any paper that uses a token to represent position instead of positional embedding?

Are you ready? 🌈 #nijijourney