Xiaodong
Xiaodong
Learning rate is too big?
@heshenghuan , I reran the script with different random seeds and didnot hit the bug as your mentioned. I'm wondering which pretrained model is used in your experiments. Yes, config...
I would like to support such task if there is a dataset. But it may take a while since I'm quite busy for other matters.
I am very busy recently and don't have a plan for tensorflow implementation now. But I'll consider it if there are many such kind of requests.
can you share me your python/pytorch version? You can direct pull my docker to run the experiment.
Thanks pointing it out!! Does this work on pytorch version 0.4x? We will test it soon.
company doesn't allow model weights released at the moment which out of my control. so sorry for this issue.
Due to security issues, the storage is not allowed to share data externally. I'm working on it...