GALIP
GALIP copied to clipboard
tobran
[CVPR2023] A faster, smaller, and better text-to-image model for large-scale training
 我在这里有看到CLIP_feat设置为true之后，在D_BLOCK中return x + self.gamma * res + self.beta * CLIP_feat， 请问CLIP_feat这个参数是怎么获得的呀，感谢！
I've structured my query_engine to output as a pydantic object: ``` class TechOffer(BaseModel): """Title of technical offer that is relevant to the query including reasons why technical offer is relevant...
Hi, how do I choose the optimal model? Do I use the last saved model directly when calculating the evaluation metrics, or do I select the best of the saved...
Hey @tobran I have read the paper and it is excellent . I implemented the paper recently . I have a doubt that can i use any other text encoder...
hi your work is great! here i wonder how to change to my dataset? please reply thanks!
Hello sir, I would like to ask if this repository has a distributed training version？
Among the parameters accepted by the NteG generator, there is a CLIP model. I want to know how to import CLIP in the GALIP code to instantiate it. I have...
Hello author! class NetD is the discriminator, right? Can I use it to replace the model of my own code? My job is image inpainting. If I change the discriminator,...
Thank you for sharing your excellent work. I would like to know something about the approximate number of epochs needed to train for the results in the paper when training...
Hi, I want to do the Super Resolution task by replace the clip text feature with the clip image feature. I think the image feature space and the text feature...
