Question about the training method and data used
Hi Phr00t! This is really amazing work — it significantly improves the speed of Qwen image editing. As a researcher, I’m very interested in how you trained it. Could you share some details about the training process, such as which training method you used and what kind of data the model was trained on?
Thanks again for your great contribution!
I don't train anything.
I merge stuff and sometimes it works.
Ah, I see! Thanks for the clarification. In that case, could you share which Qwen-Image-Edit Lightning LoRAs you merged? For example, are they from the lightx2v/Qwen-Image-Lightning repository, or from somewhere else?
Thanks again for the insight!
Just curious -- what do you research superGong? I feel like you'd understand how to know which LoRAs are in Phroot's merges just by reading the MetaData like everyone else? It's all right there.