MingTao(陶明)
MingTao(陶明)
> 我也遇到过，参数问题。 请问你后来解决了吗？
Paper title: GALIP: Generative Adversarial CLIPs for Text-to-Image Synthesis Paper link: [https://arxiv.org/abs/2301.12959](https://arxiv.org/abs/2301.12959) Code link: [https://github.com/tobran/GALIP](https://github.com/tobran/GALIP) GALIP是一个简单，快速，高质量的文本到图像生成模型，对比需要数百张GPU，400M图文对，数周时间进行预训练的Diffusion Model和Autoregressive Model，GALIP仅需8张3090，12M图文对，3天的预训练时间，取得了相当甚至更好的效果，同时生成速度提高了120倍，支持仅CPU生成。代码和预训练模型已经开源。 GALIP is a simple, fast, and high-quality **Text-to-Image Generative Model** with comparable...
Hi, we train our GALIP about 30 epochs on CC12M.
Hi, you can instantiate it as the code in the perpare_models function does. https://github.com/tobran/GALIP/blob/6a72f59dc51b1c7fd0e7671b9fac8f095d3243ed/code/lib/perpare.py#L43
> 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...
Hi, you can modify the set CUDA_VISIBLE_DEVICES=0,1 and nodes=2 in train.sh for your training. The data error looks like there is no prepared dataset under ./data. You can download it...
Hello, the coco.yml is under ./code/cfg/ folder and the npz file has been uploaded to Google Drive, you can find the link (https://drive.google.com/file/d/15Fw-gErCEArOFykW3YTnLKpRcPgI_3AB/view?usp=sharing) in the readme.
> How can I get in contact with Ming Tao? The provided email adress does not seem to work. Hello, you can contact Ming Tao ([email protected]) through e-mail.
> Hey, did you receive my mail? Hi, I have checked the inbox of [email protected], but I didn't find your mail.
> Hey, maybe it landed in your spam folder? This time I got no server rejection back. Is there another email adress I could try to contact you? Hi, I...