facebookresearch / TimeSformer

The official pytorch implementation of our paper "Is Space-Time Attention All You Need for Video Understanding?"
Other
1.51k stars 209 forks source link

Modified AVA dataset problem #99

Open MitchellWei opened 2 years ago

MitchellWei commented 2 years ago

Hi, I'm a beginner and I'm using a modified AVA dataset for some animal action recognition. I managed to run the code on SlowFast, but I found this code seems not support AVA dataset, or I missed something? I wish I could use my modified AVA dataset, so what should I do? Or, if cannot, is there any methods to transform my AVA dataset to Kinetics form so that i could use the Kinetics yaml file? Many thanks.

chenqqqqqqq commented 1 year ago

Hi, I'm a beginner and I'm using a modified AVA dataset for some animal action recognition. I managed to run the code on SlowFast, but I found this code seems not support AVA dataset, or I missed something? I wish I could use my modified AVA dataset, so what should I do? Or, if cannot, is there any methods to transform my AVA dataset to Kinetics form so that i could use the Kinetics yaml file? Many thanks. 我和你的需求一样,想要使用TimeSformer模型训练自制的AVA数据集。请问你现在有进展了吗?或许我们可以交流一下。

I have the same requirement as you and would like to use the TimeSformer model to train a custom AVA dataset. Have you made any progress on this? Perhaps we can discuss and exchange ideas.

xqc-qc commented 1 year ago

嗨,我是初学者,我正在使用修改后的 AVA 数据集进行一些动物动作识别。我设法在 SlowFast 上运行代码,但我发现这段代码似乎不支持 AVA 数据集,或者我错过了什么?我希望我可以使用修改后的 AVA 数据集,那么我该怎么办?或者,如果不能,是否有任何方法可以将我的 AVA 数据集转换为动力学形式,以便我可以使用动力学 yaml 文件?非常感谢。 我和你的需求一样,想要使用TimeSformer模型训练自制的AVA数据集。请问你现在有进展了吗?或许我们可以交流一下。

我和你有同样的要求,并希望使用 TimeSformer 模型来训练自定义 AVA 数据集。你在这方面有什么进展吗?也许我们可以讨论和交流想法。

可以交流一下,我也是自己制作AVA数据集在slowfast上运行,也想要在TimeSformer 模型来训练 I can communicate that I also made my own AVA dataset to run on slowfast, and I also want to train on the TimeSformer model