Results 12 comments of Ron Mokady

Hi @wytszhuzhu, we have used the Oscar evaluation code to evaluate over COCO I Hope this is helpful

@NielsRogge Sounds cool, I'll hope to try it out soon

> @rmokady any updates on this? Currently our top priority is to improve our model, so I guess it will be delayed If someone will find it relevant, we will...

Hi @goel-shashank, Are you using our default parameters? Did you try both the GPT-2 fine-tuning and the frozen GPT-2?

For COCO where we train both prefix and GPT-2 the loss got to 1.47 Unfortunately, the logs for the Conceptual got left on an old server and cannot access these...

Hi @goel-shashank, I find some logs for conceptual captions. This is with the resnet CLIP: ![Screen Shot 2022-02-19 at 16 33 41](https://user-images.githubusercontent.com/38734354/154805306-dde0a102-4f70-4d94-865a-6f130054b0cc.png)

This is with the Vit CLIP: ![Screen Shot 2022-02-19 at 16 35 36](https://user-images.githubusercontent.com/38734354/154805393-2c6a86db-a9d0-418c-8f0e-6652cf1d8b23.png)

You can use the pydrive downloader, similary to [this notebook](https://colab.research.google.com/github/danielroich/PTI/blob/main/notebooks/inference_playground.ipynb) (first cell) Then: `downloader.download_file("1O8OLrVNOItOJoNGMyQ8G8YRTeTYEfs0P", "/content/encoder4editing/e4e_ffhq_encode.pt")`

@BigHyf You can download val.yaml and other data files, check the vinyl download page

@BigHyf I don't find it relevant to the issue I opened, please open a new one